CN1486553A - Non-linear CDMA technique with improved detection algorithm and error correction coding - Google Patents

Abstract

A method of channel extension signal coding with enhanced error correction coding mode is provided, which is based on the nonlinear block error correction coding mode, and is realized by coding the error correction coded signal by introducing detectable patterns into the signal output before transmission. At the receiver side, the detectable pattern is decorrelated using a RAKE receiver and an enhanced decoding mode.

Description

Has the nonlinear code division multiple access technology that improves detection algorithm and error correction coding

Technical field

The present invention relates generally to the Code And Decode algorithm of communication system, be specifically related to the Code And Decode algorithm of code division multiple access (CDMA) system, can strengthen conventional error correction coding mode, the ratio character of antinoise and the peak value that transmits and average power.This technology also makes and transmits and receives intrasystem complexity and reduce to minimum.

Background technology

Up-to-date technical progress has made communication electronics faster than the speed of morning several generations electronic device, and power consumption is less and cheap.This makes the global communications market that comprises fixing and mobile device increase fast again.The continuous increase of communication technology number of users is passed through in this quick growth, and provides more service and bigger bandwidth to obtain proof to the user.Expect that this growth will continue in many years of future.

The state-of-the-art technology of multi-user comm comprises: code division multiple access (CDMA).In widely used CDMA mobile honeycomb embodiment, up to the individual signal in 64 (or 256) from base station parallel be transmitted into mobile unit.In the noise circumstance of reality, this number is subjected to the restriction of law or power control algorithm permission emission peak power, and it depends on the interference from other users.Therefore, need between the parallel C DMA excited users number of the transmit signal power of compound CDMA signal and support, carry out balance.Though the high emission signal power often causes signal reception in better covering and the receiver, this also can cause the larger interference in the neighbor cell.

The amplitude of the peak value that performance index of mobile communication system are compound CDMA signals and the ratio (PAP) of average power.High PAP is the intrinsic problem of cdma system always.People have developed shaping pulse and multiple modulation technology, and for example, the Continuous Phase Modulation technology is used to alleviate the negative effect of high PAP.High PAP also makes communication system be more vulnerable to the influence of nonlinear distortion, and nonlinear distortion is introduced by high power nonlinear amplifier in the transmitter and receiver often.

The issue table that exists among the CDMA utilizes dedicated data channel to carry out shaping pulse and multiple modulation technology now.These technology reduce the bandwidth efficiency (bandwidth efficiency is the bandwidth used of message transmission and the tolerance of the ratio of system utilized bandwidth total amount) of systems.In addition, the electronic device that number of C DMA system requirements is more expensive for example, has the big linear power amplifier of dynamic range, is used to handle signal or the encoded data channels of high PAP.Another kind of scheme is to utilize nonlinear power amplifier, but these nonlinear power amplifiers introduce nonlinear distortion for high PAP signal, may cause serious data degeneration in the receiver.

The subject matter of the current third generation (3G) wireless communication system is, under having other user's disturbed conditions during with very high data rate emission, for example, per second 1 to 2 megabit (Mbps), the limited user capacity under the given error performance.This mainly is because two factors.First factor is a spreading factor limited in the standard.Utilizing PAP is 1 orthogonal variable spreading factor (OVSF) wideband CDMA (WCDMA), and when launching with 1Mbps, the spreading factor of transmit chip is 4; When launching with 2Mbps, the spreading factor of transmit chip is 2.Second factor is suboptimum code used under the given interference environment.Select many yards CDMA (MC-CDMA) that high PAP and the robustness that lacks anti-interference and anti-nonlinear distortion are arranged.

Therefore, need to improve the performance of CDMA, allow it to operate in the third generation wireless communication system of high data rate.Any new system design not only should alleviate the problem relevant with high PAP and improve jamproof robustness, and can be in existing cdma system or the upgrading path by nature, utilizes low-cost hardware, and the embodiment of software or firmware is to obtain these raisings.

Summary of the invention

According to the present invention, code division multiple access (CDMA) technology is utilized the non-linear block code and the matrix that be used for signal encoding of a class definition for the Go-CDMA sign indicating number.Consequently code signal has healthy and strong antinoise and the character of good peak value with the ratio (PAP) of average power.In addition, the algorithm of Code And Decode Go-CDMA code signal utilizes in the Go-CDMA coded sequence intrinsic polynomial construction to improve input.

According to an alternative embodiment of the invention, propose to strengthen the Go-CDMA pattern of coding, for example comprise: convolutional encoding, Turbo coding and block encoding.These technology can be utilized the Go-CDMA decoding technique of enhancing, and this technology is utilized the polynomial construction of Go-CDMA code signal.Under the situation of convolutional encoding, the signature recognition capability of the convolutional encoding ability of new mode combinations error correction and piece Go-CDMA coding.According to embodiments of the invention, the convolutional encoding pattern is far longer than the simple ability that can obtain that is connected in series of these yards with the error correction capability of Go-CDMA coding mode combination.

Go-CDMA decoding can utilize the polynomial construction among the majority logic cataloged procedure result of Go-CDMA coding.Have by means of relevant or " intrinsic " parity error correcting encoder position of extracting during the polynomial construction decoding, can assess the cost at minimum and improve the Go-CDMA decoding down.Can also utilize this extra order to improve the decoding in multilevel coding and the decoding schema, particularly under Go-CDMA and convolutional encoding or Turbo encodes or block encoding is used in combination situation.

A kind of preferred Go-CDMA pattern that radio CDMA is communicated by letter between user and the central station comprises: in the transmitter coding pattern, convolution coder is connected in series with the Go-CDMA encoder.In receiver, first decoder stage is the soft decoder of Go-CDMA, recovers extra order from the Go-CDMA signal that receives, that is, and and parity check bit.Then, the relevant parity check bit feed-in from the soft decoding of Go-CDMA also is utilized therein by second level convolution decoder.This can finish by means of the Viterbi algorithm, and is added to the augmentation encoder of " inversion ", rather than original convolutional encoding process, and this encoder also produces this parity check bit.According to this embodiment, the effect of Go-CDMA level is the error correction capability that obtains height ratio convolution coder/decoder effectively, though there are some signal energy losses.

The existence of Go-CDMA level allows signature identification, thereby also allows to utilize ' the RAKE symbol can detect ' receiver structure, and this is impossible in pure convolution error correction coding mode.Other embodiment comprises: utilize the combination of convolutional encoding and a plurality of Go-CDMA encoders and utilize the Turbo coding or the combining of block encoding and Go-CDMA encoder, and utilize RAKE receiver.

Embodiments of the invention can be applicable to cdma communication system, provide the improvement performance that is better than conventional CDMA and tdma system in many measurements.These measurements comprise: for example, the ratio (PAP) of peak value and average power is as the error correction of l/n and n/ αHan Shuo.Herein, l is the length of Go-CDMA coded sequence, and n is that the maximum number and the α of access channel is the current number that activates access channel.This measurement also comprises: with the channel capacity of message data rate representation, as the emission error rate of signal to noise ratio (snr) function, signal-to-jamming ratio (SIR) is wherein disturbed from adjacent sub-district, the excited user number purpose upper limit in the communication cell, and the amount of calculation in the Code And Decode.The error performance that the present invention shows is far superior to prior art systems.

Embodiments of the invention are well suited in any cdma system and implement.They are particularly suitable for the high bandwidth cdma system, and for example, the third generation and Geng Gao are for cdma system.In addition, comprising mobile comm unit, the base station is in the cell site of emission parallel data message flow or any cdma system of receiving station, compare with the cdma system of routine, the Go-CDMA embodiment can allow less signal shaping expense and cheap electronic device.

First embodiment of the present invention comprises: use the Go-CDMA technology with enhancing detection, as the augmentation to known convolutional encoding.Imagination The ratio convolution coder, wherein m is the number of input information position, and n is the number of output convolutional encoding position, Go-CDMA augmentation is utilized n * lGo-CDMA sign indicating number matrix exactly, the l position of n convolution code carry-out bit as transmission of encoding after intersecting.Decoding is an inverse process, that is, utilize soft enhancing to detect, and the signal that receives of Go-CDMA decoding at first removes to intersect the signal of decoding then.Under muting situation, this recovers to enter Go-CDMA n position of decoding and the parity check bit that adds to provide the n+1 position and to estimate these positions with method for distinguishing in channel.After these first order decodings is Viterbi decoder " inversion " augmentation

Encoder, rather than original convolution Encoder, it also produces suitable parity check bit.Under muting situation, this recovers the m bit sequence and obtains the PRML estimated value with method for distinguishing from the n+1 bit sequence.Decode procedure has some losses of signal, realizes being equivalent to Convolution coder/decoder.The latter is better than the best

Ratio convolution coder/decoder.Utilize standard technique, the optimization of convolution code is significant to obtain " the best " performance, also is very direct.

Second embodiment of the present invention comprises: two or more parallel-convolutions/Go-CDMA encoding block (above first embodiment) lump together and constitute the Turbo coding structure, wherein between every pair of parallel-convolution/Go-CDMA encoding block interleaver are arranged.

The 3rd embodiment of the present invention comprises: linear or non-linear block code, rather than above convolution code.Design a kind of parity check bit of correcting from Go-CDMA decoder stage transmission error to strengthen the method for next decoder stage.

The 4th embodiment of the present invention comprises: single-stage or multistage decoding, if the intrinsic parity check digit of any Go-CDMA decoding block is not transferred to next stage, for example, last level is the Go-CDMA decoding.In this case, parity check bit is used to strengthen the detection of this a specific order, as pointed in low noise environment.The mistake of utilizing other or parity check bit to calculate replaces most probable position, can realize the detection that strengthens.

The 5th embodiment of the present invention is, replace as the orthogonal code in the MC-CDMA structure if the Go-CDMA code level (wherein carrying out majority operation) among above several embodiment is utilized the Go-CDMA sign indicating number, this is an embodiment who is referred to as MC-Go-CDMA in the reference patent.In this embodiment, before the Go-CDMA decoding, majority operation is that the signal that receives in the receiver is carried out, in order to extract the additional parity bit of using in augmentation convolution decoder or Turbo decoding or the piece decode procedure.

The method that proposes also is applied to the coding of a plurality of data-messages.This data-message can comprise: the data-message that at least one is relevant with excited users, the data-message that at least one is relevant with pseudo-excited users, and/or at least one data-message relevant with non-excited users.(excited users is to utilize communication system launching the user of information.Pseudo-excited users is the transmission that control system starts by communication channel, in order that the existence of imitation excited users.Do not carry information through the data that pseudo-excited users link sends, not by receiver decoding in the communication system yet.Non-excited users is the non-active link in the communication system, does not wherein launch or receives data or information.)

If there are a plurality of code level, the displacement level can also be arranged between every phase adjacency pair code level then.(the displacement level comprises that a coding/decoding level outputs to reversible the reordering that connects between another coding/decoding level input.)

The majority logic encoding block can be used as look-up table and implements.In the case, finish improved coding based on look-up table.

Data-message can comprise: the data element of ternary form, the data element of one pole or bipolar dual format.In addition, each data-message can be derived from the data that the gap data source receives.

The improvement that proposes is applied to the spread spectrum code division multiple access signal, comprising: the data-message stream based on the Go-CDMA sign indicating number of encoding at least, the coded message stream of scrambling is flowed and launches by communication channel in scrambling based on the encoded data message of random code.In this manner, a plurality of data-message streams can be encoded scrambling and emission by wireless medium.For example, this method can be carried out in mobile comm unit or base station.In addition, data-message stream can be correlated with, incoherent or serial data stream.When this method is when carrying out in the base station, data-message stream can be relevant with different mobile units, and each mobile unit can have a plurality of data flow relevant with it.This method can comprise: some data-message streams based on non-Go-CDMA sign indicating number of encoding at least, the non-Go-CDMA encoded data message stream and the Go-CDMA encoded data message stream of scrambling is flowed and launches by communication channel in scrambling based on the non-Go-CDMA encoded data message of random code.

A kind of method of the Go-CDMA of decoding signal can comprise: receive the coded message stream of scrambling by communication channel, descrambling coded data message flow and decoding best based on the data-message stream of Go-CDMA sign indicating number.Encoded data stream can comprise: identifying information, it is to determine according to the signal in the pilot signal or any other communication protocol process.This method can comprise: at first, the coded message that receives scrambling by communication channel flows, and any non-Go-CDMA coded data message flow of descrambling separates non-Go-CDMA encoded data message stream and Go-CDMA encoded data message stream based on identifying information.Next, decode dividually non-Go-CDMA encoded data message stream and Go-CDMA encoded data message stream.

Description of drawings

With reference to detailed description and drawings, can understand above-mentioned feature and advantage of the present invention more fully, wherein:

Fig. 1 represents to have the communication channel of additive noise.

Fig. 2 represents multi-address coding-the separate code communication system according to the embodiment of the invention.

Fig. 3 A represents the functional-block diagram of used cdma system in the mobile communication, comprising the Code And Decode piece according to the embodiment of the invention.

Fig. 3 B represents that a plurality of mobile units carry out the schematic diagram of cellular communication by make an uproar wireless channel and base station are arranged.

Fig. 4 A represents the functional-block diagram of used Go-CDMA system in the mobile communication, and its expression is used at random scrambled code to the Code And Decode pattern according to the embodiment of the invention.

Fig. 4 B represents the functional-block diagram of used Go-CDMA system in the mobile communication, its expression is used scrambled code at random to Code And Decode pattern that allows the Go-CDMA coding and the another kind of coding mode implemented simultaneously in same base or mobile unit, perhaps, according to the embodiment of the invention, can in the bimodulus mobile unit configuration of mobile unit, implement.

Fig. 5 presentation code ratio is

With constraint length be the general convolution coder of Mm.

Fig. 6 represents standard ratio Convolution coder.

Fig. 7 represents the basic block diagram of Turbo encoder.

Fig. 8 represents the example of eight state recursive system convolution coders.

Fig. 9 represents the basic block diagram of two stage Turbo decoders.

Figure 10 represents the grid representation of finite state encoder.

Figure 11 represents constant lattice figure of the time of convolution code.

Figure 12 represents variable lattice figure of the time of block code.

Figure 13 represents to receive by frequency-selective channel the best RAKE receiver structure of broadband binary signal.

Figure 14 represents signal amplitude coefficient ρ ₁And the relation between the total number of channels n.

Figure 15 represents ratio with solid line Convolution coder and dot follow-up Go-CDMA coding and " soft " strengthens the effect of Go-CDMA decoding, it becomes " quite " height ratio to this coder transitions

The convolutional encoding system.

Figure 16 A represents that the Go-CDMA nonlinear code in the communication system transmit-receive letter machine divides multi-address coding and decoding and convolution error coding and decoding and intersection and remove to intersect the functional-block diagram of embodiment.

The variation of embodiment shown in Figure 16 B-16D presentation graphs 16A.

Figure 17 represents the embodiment of secondary convolution/Go-CDMA Code And Decode channel expansion pattern embodiment.

Figure 18 is non-overlapped embodiment, wherein represents ratio with solid line

Convolution coder and dotted line represent that follow-up Go-CDMA coding and " soft " strengthen the effect of Go-CDMA decoding, and it becomes " quite " height ratio to this coder transitions

The convolutional encoding system.

Figure 19 is overlapping embodiment, wherein represents ratio with solid line Convolution coder and dotted line represent that follow-up Go-CDMA coding and " soft " strengthen the effect of Go-CDMA decoding, and it becomes " quite " height ratio to this coder transitions The convolutional encoding system.

Figure 20 represents to comprise channel error correcting encoder (330a) and decoding (330b) in the communication system, channel expansion (210) and remove the functional-block diagram of expansion (220) and modulated forward conversion (211) and reciprocal transformation (221), this communication system is an actual coding device of the present invention.

Embodiment

I. summarize

The up-to-date technology of multi-user comm comprises: CDMA and time division multiple access (TDMA).These technology are widely used in single sub-district or many community mobile communications, and wherein TDMA is the basis of gsm mobile telephone system.

Fig. 1 represents the environment that multi-user comm exists.With reference to Fig. 1, communication channel 100 is expressed as the communication channel of additive noise.As an example, communication channel can be an air, the space, and such as lead, transmission line, or the electrical connection of microwave component, or optical fiber.Transmission produces signal (s+noise) by the influence that the incoming signal s of communication channel 100 is subjected to additive noise in the communication channel at the transmission line far-end.

Fig. 2 represents multi-address coding-the separate schematic diagram of code communication system 200.Communication system 200 comprises as a plurality of message transfers that are input to encoding block 210.Encoding block 210 these message of coding and by the communication channel 100 launching code message of making an uproar as composite signal.Decoding block 220 receives the composite signal that comprises noise and by such process this coded message of decoding, under the situation of zero interchannel noise, this process generally is the inverse process of cataloged procedure, or approaches inverse process at least.

Encoding block 210 and decoding block 220 all are expressed as monolithic.For example, under the situation of multiplexing fiber-optic communication system, may be single encoded and decoding block really, they are connected to the optical fiber as communication channel 100.Perhaps, in mobile communication system, one or two in encoding block 210 and the decoding block 220 can be the encoding block or the decoding block of uncoupling, and each user has one, and unique relative tertiary location is arranged between every.Fig. 3 A and 3B represent this situation.

Fig. 3 A represents a communication system.This communication system can comprise: for example, mobile unit 300 or mobile comm unit 310 are as employed in the cellular communication.System 300,310 can comprise: be coupled to the modulation/demodulation unit 320 of antenna 340, coding unit 210 and decoding unit 220, optional preposition and rearmounted coding and decoding unit 330, processor 350, memory 360 and I/O unit 370.

Processor 350 can be a microprocessor, microcontroller, digital signal processor, application-specific integrated circuit (ASIC), or any other devices that are suitable for control system 300,310 operations.Processor 350 control device 300 and 310 operation, and can be coupled to each functional block of their operations of control.Perhaps, any or all functional block can be implemented on processor in the device 300,310.By program stored instruction in the execute store 360, processor can control device 300,310, makes functional unit become activation, and is irrelevant with their specific embodiment.

Memory 360 storage data also can storage device 300,310 inner treaters 350 or the program command carried out of other unit.Memory can comprise: volatile memory, and nonvolatile memory, or the two.Memory can comprise: for example, read-only memory (ROM) and ROM device, as the CD-ROM device, hard disk and floppy disk, random-access memory (ram), database and any other type memory or storage arrangement.

I/O unit 370 can comprise the input/output device of any kind.These devices can comprise: display, and keyboard, microphone, loudspeaker, video camera, vibrating device is connected to the modem of network, and as PSTN, the interference networks of local area network (LAN) or wide area network or server are referred to as the router and the bridge of the Internet.

At run duration, processor 350 can make device 300,310 open communication channel with another communicator through antenna 340, and this communicator is followed CDMA or TDMA agreement.Under the situation of wireless cellular telephony, communication channel can be used for carrying out call.Processor 350 can also be from I/O unit 370 or memory 360 received signals, for example, speech or data-signal, and based on the data or the voice signal that receive, can dateout message to preposition and rearmounted coding and decoding unit 330.Data-message can send according to this by coding unit 210 and modulation/demodulation unit 320, and from following the antenna output of suitable communication protocols.Similarly, in the opposite direction, processor can be through antenna 340, and decoding unit 220 and preposition and rearmounted decoding unit 330 receive data-message.Then, based on the data-message that receives, processor can output signal or other data to one or more I/O unit 370, or the storage data are to memory 360.In this manner, on behalf of this device users, device 300,310 can finish communication function.

Preposition and rearmounted coding and decoding piece 330 is chosen wantonly.Its purposes can be to insert (being deciphering) error correcting code under the decoding situation in data-message, or before coding or after the decoding, intersects or goes to intersect data or deal with data message.Under the situation of inserting error correcting code, can utilize any error correction or error protection pattern, comprise Cyclic Redundancy Check pattern and forward error correction (FEC) pattern.

The Code And Decode piece can be conventional CDMA or TDMA encoding block, or the title of October 5 calendar year 2001 application is the Go-CDMA encoding block described in the U.S. Patent application of " Method and Apparatus for Non-LinearCode-Division Multiple Access Technology ", it is incorporated in this is for reference.Encoding block 210 and decoding block 220 can be implemented the Go-CDMA Code And Decode pattern according to embodiments of the invention.Perhaps, Go-CDMA coding and CDMA coding can be present in the hybrid system shown in Fig. 4 B simultaneously.

Modulation/demodulation unit 320 can combine with any suitable amplifier, is used to set up the modulated output signal s based on TDMA or CDMA pattern, comprises the CDMA pattern of utilizing the Go-CDMA technology.

Consider CDMA mode communication apparatus 300,310, it comprises: the processor of execution of program instructions or other devices are used to finish the decoding function of the encoding function and the decoding block 220 of encoding block 210.In the case, can utilize data and program command updated stored device 360, configuration codes piece 210 and decoding block 220 are to implement according to Go-CDMA Code And Decode pattern of the present invention.Program command and data or the data that receive through antenna 340 can be encased in memory 360 through one or more I/O unit 370.

Fig. 3 B represents that a plurality of mobile units 310 carry out the schematic diagram of cellular communication by make an uproar wireless channel 100 and base station 300 are arranged.Mobile radio station 300 and each coding unit 210 separately can be considered to and be equivalent to single encoded unit 210 shown in Figure 2, and n the data message that is used to encode is launched by noisy channel 100.In this case, base station and decoding unit 220 thereof can be considered to and be equivalent to single decoding unit 220 shown in Figure 2, and n of the composite signal plus noise that is used to decode receives data-message.

In the wide-band mobile communication of third generation standard, chosen the CDMA scheme.Potential technology is the majority logic coding, and it does not also send any widely used communication system in a large number to.This potential technology is the basis of Go-CDMA technology of the present invention in the co-applications patent, and therefore, it also is the basis that the present invention strengthens the Go-CDMA technology.

The polynomial construction of majority logic code signal

Go-CDMA nonlinear code division multiple access technology partly utilizes sign or sgn majority logic function, wherein:

With

Or (1.2)

The majority logic code signal, for example, Go-CDMA code signal s (τ) | τ ∈ [0, t] is as the function of time τ, according to as give a definition:

$s\left(\mathrm{\τ}\right)=\mathrm{sign}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{d}_i{X}_i\left(\mathrm{\τ}\right)---\left(1.3\right)$

Or

$s_{\±}\left(\mathrm{\τ}\right)={\mathrm{sgn}}_{\±}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{d}_i{X}_i\left(\mathrm{\τ}\right)---\left(1.4\right)$

I=1 wherein ..., n channel number, d _i± 1 is the information bit of i channel, and X _i(τ) | τ ∈ [0, t] is the bipolar code word of i channel.Code word is in time cycle [0, t] interior m interim equal time piecewise constant.In some applications, it is orthonormal requiring code word: $\underset{0}{\overset{1}{\∫}}{X}_j\left(\mathrm{\τ}\right)X_i\left(\mathrm{\τ}\right)\mathrm{d\τ}={\mathrm{\δ}}_{\mathrm{ij}},$ I wherein, j=1 ..., if n is i ≠ j, then δ _Ij=0; Otherwise δ _Ij=1, but always not this situation.

In 1964, following s (τ) is proposed | the multinomial series structure of τ ∈ [0, t]:

$s\left(\mathrm{\τ}\right)={\mathrm{\ρ}}_0+{\mathrm{\ρ}}_1\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{d}_i{X}_i\left(\mathrm{\τ}\right)+{\mathrm{\ρ}}_2\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{j=i+1}{\overset{n}{\mathrm{\Σ}}}{d}_i{d}_j{X}_i\left(\mathrm{\τ}\right)X_j\left(\mathrm{\τ}\right)+\·\·\·---\left(1.5\right)$

$\&CenterDot;\&CenterDot;\&CenterDot;+{\mathrm{\&rho;}}_n{d}_1{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">d</figure-callout>}}_2\&CenterDot;\&CenterDot;\&CenterDot;d_n{X}_1\left(\mathrm{\&tau;}\right)X_2\left(\mathrm{\&tau;}\right)\&CenterDot;\&CenterDot;\&CenterDot;X_n\left(\mathrm{\&tau;}\right),$

The formula that wherein provides is only to calculate ρ according to n ₀, ρ ₁And ρ _n, as follows:

ρ ₀＝0

For i=2,3 ..., n-1, | ρ _i|＜ρ ₁

We observe herein, and all possible selection provides (n+1) individual unknown number ρ ₀, ρ ₁... ρ _n(n+1) individual linear formula, and, can find the solution all ρ for the integer of any n 〉=2 _iEqually, when n is even number, for the different s of s (τ) _±(τ), identical polynomial construction is arranged, and can calculate unknown number ρ ₀, ρ ₁... ρ _nIn fact, when n is even number, the ρ among the s (τ) ₀=0, and s _±ρ (τ) ₀≠ 0.When n is odd number, s _±(τ)=and s (τ), at all unknown number ρ ₀, ρ ₁, ρ ₂... ρ _nIn, ρ ₀=ρ ₂=...=ρ _N-1, therefore, s _±Coefficient ρ (τ) _iBe ρ ₁, ρ ₁, ρ ₃, ρ ₃... ρ _N-2, ρ _N-2, ρ _n, wherein (n+1) is even number.

Equally in some applications, require these functions X _i(τ), X _i(τ) X _i(τ) ..., X ₁(τ) X ₂(τ) ... X _n(τ) be mutually orthogonal.

The best PRML of majority logic code signal detects

When the data flow of n channel of coding, because data have 2 ⁿIndividual different combination, so 2 ⁿIndividual different signal s _*(τ) | τ ∈ [0, t] also is possible.Herein, s _*(τ) expression s (τ) or s _±(τ).Therefore, measure, utilize 2 according to the least square of standard ⁿIndividual correlator or matched filter banks, 2 ⁿEach signal in the individual possibility signal has one can realize the optimum detection of majority logic to choose the signal that optimum Match is arranged, particularly the n position signal data vector { d of Go-CDMA coding ₁, d ₂... d ₃.The cost of this optimum detection algorithm is the complexity of detection architecture, and it is to increase by exponential manner with data channel number n.

The conventional suboptimum of majority logic code signal detects

Conventional majority logic decoding schema is based on following rule estimation d _iInformation:

${\hat{d}}_i=\mathrm{sign}\left(z_i\right);$ $z_i=\underset{0}{\overset{t}{\&Integral;}}{s}_{*}{X}_i\left(\mathrm{\&tau;}\right)\mathrm{d\&tau;},$ I=1 wherein ... n.(1.7)

Can utilize this detection algorithm with decoding Go-CDMA nonlinear code division multiple access technology.According to formula (1.7) and (1.3), can observe this detection algorithm and under required orthogonality, extract nonzero term in the polynomial construction (1.5).This is effectively, because the nonzero term that can observe among the polynomial construction s (t) according to formula (1.6) is carried most information bit signal energy.

The improvement suboptimum of majority logic code signal detects

At n is under the situation of odd number, and the actual augmentation that above suboptimum detects is the last term of extracting in decoding among the polynomial construction s (τ), because this also carries d _iMost information position signal, as follows:

(a) estimation z _iCorrelation, i=1 wherein ..., n+1,

$z_i=\underset{0}{\overset{t}{\&Integral;}}s\left(\mathrm{\&tau;}\right)X_i\left(\mathrm{\&tau;}\right)\mathrm{d\&tau;},$ i＝1，...，n，

$z_{n+1}=\underset{0}{\overset{t}{\&Integral;}}s\left(\mathrm{\&tau;}\right)X_1\left(\mathrm{\&tau;}\right)X_2\left(\mathrm{\&tau;}\right)\&CenterDot;\&CenterDot;\&CenterDot;X_n\left(\mathrm{\&tau;}\right)\mathrm{d\&tau;}---\left(1.8\right)$

(b) estimation information bit

Defer to following rule:

${\hat{d}}_i=\mathrm{sign}\left(z_i\right)\mathrm{if}\left|z_i\right|>\left|z_{n+1}\right|,$ I=1 wherein ..., n (1.9)

Perhaps, as | z _i|＞| z _j| the time, i=1 wherein ... j-1, j+1 ... n+1, and j=1 ..., n.

${\hat{d}}_j=\frac{\mathrm{sign}\left(z_{n+1}\right)\mathrm{sign}\left({\mathrm{\&rho;}}_n\right)}{{\hat{d}}_1{\hat{d}}_2\&CenterDot;\&CenterDot;\&CenterDot;{\hat{d}}_{j-1}{\hat{d}}_{j+1}\&CenterDot;\&CenterDot;\&CenterDot;{\hat{d}}_n}---\left(1.10\right)$

This improvement can produce the better estimated value of a n information bit, and this estimation that is based on correlation in (1.5) last term replaces the most weak information bit estimated value Be this situation under the low noise that meets the following conditions:

(a) mistake occurs in the point of least correlativing coefficient value, | z _j|;

(b) every other information bit Be zero defect, i=1 wherein ... n and i ≠ j.

Herein, we observe, and it is odd number and only consider s that identical method can be used for situation that n is an even number and n _±(τ) situation of replacement s (τ).

The Go-CDMA sign indicating number

These sign indicating numbers be October 5 calendar year 2001 day header for described in the co-applications United States Patent (USP) of " Method and Apparatusfor Non-Linear Code-Division Multiple Access Technology ", it is incorporated in this is for reference.Put it briefly, these sign indicating numbers are elements set 1, the sign indicating number of generator matrix G form among the+1}, for element gather 1, the bipolar vector d the among+1}, then $\hat{d}:={\mathrm{sgn}}_{\&PlusMinus;}\&lsqb;G^T{\mathrm{sgn}}_{\&PlusMinus;}\left(\mathrm{Gd}\right)\&rsqb;=d,$ If d is a ternary, or in the particular subset of trinary data is closed, then $d^T|\hat{d}-d|=0,$ Wherein | x| is the amplitude of x.This means, in the building block of multi-address communication system, form the bipolar data of excited users of data d vector and possible pseudo-excited users data and zero non-excited users " data " and can be decoded into s=sgn _±(Gd).For the error-free communication in the zero interchannel noise incident, this amount is launched and is decoded into

$\hat{d}:={\mathrm{sgn}}_{\&PlusMinus;}\&lsqb;G^{T}s{\mathrm{gn}}_{\&PlusMinus;}(s+\mathrm{channel\; noise})\&rsqb;---\left(1.11\right)$

Perhaps, it by chance is healthy and strong estimated value.

Multistage Go-CDMA coding

Parallel encoding or decoding block constitute coding or decoder stage, are connected in series these grades formation multilevel coding and decode system.Intergrade can relate to various in-place computations.

Convolutional encoding

According to embodiments of the invention, convolutional encoding and decoding technique can be used for the coding transmission signal.An example of convolution decoder technology is the PRML decoding that is referred to as the Viterbi algorithm.

Armed information sequence transmission can realize convolutional encoding by linear finite state register.In general, with reference to figure 5, shift register 504 comprises M (m position) level and n linear algebra function generator 503.The input data 501 of supposing encoder 500 are binary datas, and displacement enters shift register 504 according to this, each m position.The carry-out bit number of each m position list entries is the n position.Therefore, code check is defined as Parameter M is referred to as the constraint length of convolution code.

A kind of method of describing convolution code is to utilize its generator matrix.In general, the generator matrix of convolution code is semi-infinite, because the length of list entries is semi-infinite.As the another kind of scheme of regulation generator matrix, we stipulate limited dimension generation system.We stipulate one group of n vector, give vector of each adder in n modulus 2 adders 503.The amplitude of each vector is Mm, and comprises the connection of encoder 500 to modulus 2 adders 503.

In particular, we consider binary convolution encoder 500, its constraint length M=3, and m=1 and n=3, Fig. 6 represent this situation.At first, suppose that shift register 504 is in all-zero state.Suppose that the first input position is ' 1 '.Then 3 output sequence is ' 111 '.Suppose that second is ' 0 ', then output sequence is ' 001 '.If the 3rd is ' 1 ', then output is ' 100 ', or the like.Now, suppose that we are numbered 1,2 and 3 for the function maker output that produces per 3 output sequences, from top to bottom, similarly, give each respective function maker numbering.So because only the first order is connected to the first function maker (not needing modulus 2 adders), then maker is

g ₁＝[100] (1.12)

The second function maker is connected to the first order and the third level.Therefore.

g ₂＝[101] (1.13)

At last,

g ₃＝[111] (1.14)

Sometimes, the maker of planting sign indicating number is to provide with the octal system form easily, as (4,5,7).

Different with the block code that regular length n is arranged, convolution coder is finite state machine basically.So the best decoding of convolutional encoding sequence relates to dynamic programming and searches for by grid to obtain the most probable sequence.Detector carries out hard decision or soft decision after depending on demodulator, and the correspondence metric in the grid search is respectively Hamming tolerance or Euclidean tolerance.The best decoder of convolutional encoding is referred to as the Viterbi algorithm, and it has PRML sequence estimation amount.

The details that in application text, provide the present invention to utilize convolutional encoding, but in standard textbook, can find standard convolutional encoding and decoding, the definition of different sign indicating number tolerance and the more particulars of best Viterbi algorithm.According to the present invention, can utilize any convolutional encoding and decoding schema.

The Turbo coding

The citation form of the encoder of Turbo encoder 400 comprises: two systematic encoders 410 that link together by means of interleaver 420, as shown in Figure 7.The Turbo code utilization is only to the pseudorandom interleaver 420 of system bits computing.In the Turbo coding, utilize interleaver that two reasons are arranged, that is:

Half incidental mistake in the Turbo code and in second half utmost point do not allow incident mistake to link together.This really why the Turbo coding because the main cause of conventional code.

Provide healthy and strong performance to mismatch decoding, the mismatch decoding is a problem unknown in channel statistical or that occur during by wrong regulation.

Usually, but needn't be like this, identical sign indicating number is used for two encoders 410 of Fig. 7.Composition sign indicating number to the Turbo code suggestion is short constraint length recursive system convolution (RSC) sign indicating number.Provide the example of eight state rsc encoders 410 among Fig. 8.The reason that makes convolution code become recurrence (input is got back in one or more tap output feed-ins in the shift register) is to make the internal state of shift register depend on output in the past.The feature (single error in the system bits produces the parity error of infinite number) that this influences the mistake figure consequently realizes the better performance of overall coding strategy.We also know, utilize the conversion of particular transform algorithm, and each rsc encoder 410 has suitable standard convolution coder.

In Fig. 7, input directly is added to encoder 1, and the pseudorandom of the same data stream version that reorders is added to encoder 2.Two group coding positions of system bits (that is original message bits) and two encoder 410 generations constitute the output of Turbo encoder 400.Though becoming demal is convolution code, in fact, Turbo code is a block code, and its block length is determined by interleaver 420 length.In addition, because two rsc encoders 410 among Fig. 7 are linear, we can describe Turbo code is linear block code.

In receiver, the signal that two stage Turbo encoders 450 shown in Figure 9 can be used for encoding and receive.In described example, every grade in two decoder stage 451 is utilized Bahl, Cocke, and Jelinek and Raviv (BCJR) algorithm is to solve (MAP) detection problem of maximum " posterior probability ".Bcjr algorithm is different from the Viterbi algorithm at following two basic sides:

1.BCJR algorithm is " soft input, soft output " decoding algorithm that twice recurrence arranged, forward direction and once the back to, this twice recurrence all relates to soft decision.Contrast therewith, Viterbi algorithm are " soft input, hard output " decoding algorithms, and the single forward recursive relates to soft decision; The recurrence end is hard decision, therefore obtains specific remaining path between several other paths.Consider that from the evaluation work aspect because backward recursive, bcjr algorithm is more complicated than Viterbi algorithm.

2.BCJR algorithm is the MAP decoder, by " posterior probability " of individual bits in the estimation code word, it makes the faults minimum; In order to rebuild original data sequence, the soft output of bcjr algorithm is hard-limiting.On the other hand, the Viterbi algorithm is a PRML sequence estimation amount, and it makes whole sequence rather than every 's likelihood function maximum.Therefore, bcjr algorithm can be better than the Viterbi algorithm slightly; It will not be poorer.

In Turbo decoder shown in Figure 9, the system bits u ' of reception _kWith bits of coded d ' ¹ _kFeedback data with the early decoding process is handled by first order BCJR decoder 451.The result of this process is intersected by pseudorandom interleaver 420 again.The output of interleaver and the bits of coded d ' of reception ² _kHandled by second level BCJR decoder 451 together.The output of this second level BCJR decoder 451 is gone interleaver 421 to go to intersect by two pseudorandoms of separating.A pseudorandom goes the output of interleaver to feed back to first order BCJR decoder, as receiving position decode procedure afterwards.Another pseudorandom goes the output transmission of interleaver by hard limiter threshode devices 430, obtains transmitting the estimated value of information bit.

The details that in application text, provide the present invention to utilize Turbo coding, but in standard textbook, can find the more particulars of standard Turbo Code And Decode and best bcjr algorithm.Should be understood that according to the present invention and can utilize any Turbo Code And Decode pattern.

The linear block coding

The normal linearity coding techniques, for example, the sign indicating number generator matrix, parity-check matrix and bug list can be as the part cataloged procedures according to the embodiment of the invention.

Block code comprises one group of regular length vector that is referred to as code word.Code word size is the element number in the vector and is masked as n.Element in the code word is selected from the letter of q element.When letter was made up of two elements 0 and 1, this sign indicating number was a binary code.When code word element is selected from q element alphabetical (q＞2), this sign indicating number is a nonbinary code.In being the binary system block code of n, length has 2 ⁿIndividual possibility code word.From these 2 ⁿIn the individual code word, we can choose V=2 ^mIndividual code word (m＜n) to form a sign indicating number.Therefore, the piece of m information bit is transformed into and is selected from V=2 ^mLength is the code word of n in the individual codeword set.The block code ratio that forms is

Make u _V1, u _V2... u _VmThe presentation code vector code word N that embarks on journey _vM information bit.Therefore, m the information bit sign vector that enters the linear block code encoder is

U _v＝[u _v1，u _v2，...u _vm] (1.15)

And the output of encoder is vector

N _v＝[n _v1，n _v2，...n _vn] (1.16)

The coding computing of finishing in the linear binary block encoder can be represented with one group of n equation form.

n _vj＝u _v1g _1j+u _v2g _2j+...+u _vmg _mj，j＝1，2，...n (1.17)

G wherein _Ij∈ 0,1}, and u _Xig _IjRepresent u _XiWith g _IjProduct.These formula can also be expressed as matrix form:

N _v＝U _vG (1.18)

Wherein G is referred to as the generator matrix of sign indicating number, and G is expressed as

$G=\left[\begin{array}{cccc}{g}_{11}& g_{12}& \&CenterDot;\&CenterDot;\&CenterDot;& g_{1n}\\ g_{21}& g_{22}& \&CenterDot;\&CenterDot;\&CenterDot;& g_{2n}\\ \underset{.}{\overset{.}{.}}& \underset{.}{\overset{.}{.}}& \underset{.}{\overset{.}{.}}& \underset{.}{\overset{.}{.}}\\ g_{m1}& {\mathrm{<figure-callout\; id="2"\; label="g\; m"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">g</figure-callout>}}_m& \&CenterDot;\&CenterDot;\&CenterDot;& g_{\mathrm{mn}}\end{array}\right]----\left(1.19\right)$

Can be by row operation (and column permutation) ratio The generator matrix of linear block code is reduced to " system form "

I wherein _mBe m * m unit matrix, and P is the matrix of the m that determines n-m redundant digit or parity check bit * (n-m).For the system linear block code, its generator matrix G is provided by its " system form ", and then parity check matrix H is

H＝[-P′|I _n-m] (1.21)

Wherein P ' is a P transpose of a matrix matrix.When handling binary code, the negative sign the among-P ' can omit, because modulus 2 subtractions are identical with modulus 2 additions.

Generator matrix G is used in the coding computing in the transmitter.On the other hand, parity matrix is used in the decoding computing in the receiver.Under the linguistic context of latter's computing, make r represent that 1 * n receives vector, it is to obtain by the code vector c that noisy channel sends.Therefore, r can be expressed as following form

r＝c+e (1.22)

Wherein e is referred to as error vector or error mode.If the corresponding element of r is identical with the element of c, then the i element of e equals ' 0 '.On the other hand, if the corresponding element of r is different with the element of c, then the i element of e equals ' 1 ', in the case, just says that error occurs in the i position.The task of receiver is according to receiving vector r decoding code vector c.Being generally used for finishing this algorithm of decoding computing is at first to calculate 1 * (n-m) vector that is referred to as fault.Provide 1 * n and receive vector r, corresponding fault definition is

Fault=rH ' (1.23)

Wherein H ' is the matrix transpose of parity check matrix H.The complete form of fault is referred to as bug list usually.

The details that in application text, provide the present invention to utilize linear block coding, but in standard textbook, can find the more particulars of calibrated bolck Code And Decode.Should be understood that according to embodiments of the invention, can utilize any linear block coding mode.

Grid coding

The sign indicating number lattice are diagrammatic representations of convolution code or block code, and wherein every paths is represented code word (or sign indicating number sequence).This expression might be implemented the PRML decoding (MLD) of sign indicating number, can reduce the complexity of decoding greatly.The grid base MLD algorithm of knowing most is the Viterbi algorithm.The error control of convolution code grid representation in the Viterbi decoding algorithm causes convolution code over the past two decades Applied Digital is communicated by letter widely.The grid representation of linear block code also causes effective MLD pattern of linear block code.

The Coded by Linear Codes device that comprises the convolution code and the block code of finite memory, the model that can be used as finite state machine, wherein Γ={ 0,1, any instantaneous output code position of 2...} is by current input information position and coder state is definite uniquely at this moment during encoded interval.Utilize this model, the dynamic characteristic of encoder represents it is lattice diagram form shown in Figure 10.The trellis encoder of convolutional encoding process has mesh transformations, and it begins and proceed to next state from nought state, and is relevant with system next input.Usually, for given system restriction length, preceding after what, mesh transformations will arrive the transient state that repeats with state in the past.Be typically, trellis encoder is to be made of finite state encoder and transform block, and transform block becomes signal in the given signal constellation which to the output transform of encoder.

Usually, grid decoder is to constitute (for example, Viterbi decoder) by one group of filter and a MLD decoder.Signal Matching in each filter and the given planisphere.The output of filter is used to calculate the branch metric of Euclidean range measurement.This tolerance is used for the MLD decoder to determine the PRML list entries.

The lattice figure of convolution code generally is that the time is constant.Yet, the lattice figure of linear block code normally the time variable.Figure 11 and 12 represents the variable lattice figure of time of constant lattice figure of the time of convolution code and linear block code respectively.

The details of the uniform enconding grid representation of convolutional encoding of the present invention and linear block coding are provided in application text, but in standard textbook, can find standard network trellis coding and decoding, different trellis code defined matrixs and the optimal mapping sign indicating number lattice more particulars on the modulated signal constellation.According to embodiments of the invention, can utilize any grid coding and decoding schema.

Utilize convolution or the Turbo or the block encoding of trellis coded modulation

Output in convolution or Turbo or the background block error correcting encoder process can transform to maximum Euclidean distance or maximum Hamming distance from a networking trellis code word.Finish trellis coded modulation (TCM) technology that this conversion can propose in nineteen eighty-two based on Ungerboeck or according to the deriving technology of this technology.

The optimal detector of this TCM based system is PRML detector (MLD).MLD base detector often has complicated hardware, and complexity increases by exponential manner with the signal sum that may detect.

According to embodiments of the invention, can utilize any trellis coded modulation and demodulation modes.One of this TCM basic mode formula is that the output transform of the embodiment of the invention is become non-linear block code matrix.

The nonlinear code matrix

The nonlinear code matrix is a rectangle nonlinear code matrix, and if possible, this matrix is to be made of the less non-linear and linear code matrix that satisfies the Plotkin boundary.This non-linear code word matrix utilizes the Levenshtein structure to constitute.If find that non-linear code word matrix not in the Plotkin boundary, then utilizes other building method, for example, utilize | u|u v| building method.Because the non-linear code word matrix of the latter is not in the Plotkin boundary, their concrete length does not have best minimum code distance.Yet the minimum range of these non-linear code word matrixes can have maximum probable value.

At given minimum code distance c _Min-distDown, the Plotkin boundary is the possible maximum length upper limit of the non-linear code word matrix of control.

The Plotkin boundary

If c _Min-distBe even number and 2c _Min-dist＞l, then

$m\&le;2\&lsqb;\frac{c_{\min -\mathrm{dist}}}{{2c}_{\min -\mathrm{dist}}-l}\&rsqb;,$

Wherein l is the length of nonlinear code word, and 2 ^mIt is possible non-linear number of codewords.If c _Min-distBe odd number and 2c _Min-dist+ 1＞l, then

$m\&le;2\&lsqb;\frac{c_{\min -\mathrm{<figure-callout\; id="1"\; label="dist\; +"\; filenames="A0182193800571.png,A0182193800631.png"\; state="\{\{state\}\}">dist</figure-callout>}}+1}{{2c}_{\min -\mathrm{dist}}+1-l}\&rsqb;.$

The Plotkin boundary is the good measure that the given length of non-linear block code descends best minimum code distance.

The Levenshtein structure

This building method can utilize the Hamming matrix of smaller length.In order to construct nonlinear code ξ, just require following definition:

1. if 2c _Min-dist＞l 〉=c _Min-dist, then definition $k=\frac{c_{\min -\mathrm{dist}}}{{2c}_{\min -\mathrm{dist}}-l},$ With

α＝c _min-dist(2k+1)-l(k+1)，β＝kl-c _min-dist(2k-1)。

α and β all be nonnegative integer and

l＝(2k-1)α+(2k+1)β

c _min-dist＝kα+(k+1)β

Therefore, when l was even number, then α and β were even numbers.When l is odd number and k when being even number, then β is an even number, otherwise when l and k were odd number, then α was an even number.

Utilize the Levenshtein structure, nonlinear code ξ is:

Wherein is modulus 2 additions, A _nBe deletion first row and length be n binary system Hadamard (all elements be gather 0, among the 1}), and A ' _nBe at A _nThe code word that basis zero begins in the matrix is also deleted the zero initial set of matrices of deriving.

| u|u v| structure

This building method is to set up non-linear code word matrix in such a way from less matrix:

1. suppose (l, 2 ^M1, c _Min-dist1) sign indicating number ξ ₁(l, 2 ^M2, c _Min-dist2) sign indicating number ξ ₂Identical length is arranged, can form new nonlinear code ξ ₃, comprise all vectors

ξ ₃＝|u|uv|，u∈ξ ₁，v∈ξ ₂，

Wherein all non-linear code words be binary system 0,1}

2. the sign indicating number ξ that forms ₃Be (2l, 2 ^M1m2, c _Min-dist3=min{2 c _Min-dist1, c _Min-dist2)

In conjunction with above various building methods, can extract the nonlinear code matrixing of the present invention's output from the row and column of quadrature Hadamard matrix, or according to less non-linear code word matrix construction, or from the output signal of itself, derive.

The RAKE receiver structure

When the processing propagation was selected the signal of broad-band channel by frequency, the problem of multipath fading often appearred.If L different mulitpath is to add up independently, then we can obtain L duplicate of identical traffic signal in receiver.Therefore, press best mode and handle the performance that the receiver of received signal obtains the L level diversity communication system of equivalence.RAKE receiver is the receiver that diversity is optimized.Figure 13 represents to utilize the RAKE receiver ideal device of wall scroll delay line, and signal r (t) transmission of reception is by this delay line.The signal of each tap place is and X _Nj(t) pn _Ch(t) relevant, j=1 wherein, 2 ... L and ch=1,2.Figure 13 represents the structure 700 of this receiver.In fact, tapped delay receiver 700 is attempted to collect signal energy from all signal paths, and these signal paths drop in the span of delay line and carry identical information.In Figure 13, every is detected the path by each tap place

Postpone 701 separately.In in-phase channel (i=1), also be referred to as real channel and, also be referred to as empty phase channel at quadrature phase channels (i=2), the broadband signal of reception is at first by pseudo random sequence pn _i702 descramblings.Be signaling interval X subsequently for different chip offset _n(t), the signal that receives on the every paths of decorrelation.In Figure 13, this is in step 703, finishes in 704 and 705.In addition, in step 704, the signal plus of all RX path makes the signal strength signal intensity maximization of reception together.At last, step 706 is removed the phase difference between reality bit channel and the empty phase channel, and the information of detection is fed into sample-and-hold device 707, prepares to do further decoding.

Be important to note that,, then can utilize the matched-filter approach among Figure 13 to finish " RAKE symbol detection " if the signal that receives has known signature or symbol, otherwise, " detection of RAKE chip " method only can be attempted.Last method is easier to implement, and can do very effectively, and back one method requires very complicated chip signal follower, and is easy to generate intersymbol interference.

The details that in application text, provide the present invention to utilize the RAKE receiver structure, but in standard textbook, can find the more particulars of different RAKE receiver structures.Should be understood that according to embodiments of the invention, can implement any RAKE receiver pattern.

Intersect

Interleaver is the input-output converting means, replaces the ordering of symbol sebolic addressing in the fixing letter according to complete certainty mode, in other words, takes out symbol at input, and produces the same-sign of different time order at output.Therefore, interleaver is effective ways of handling sudden mistake channel, because its intersect coding data in such a manner makes sudden channel conversion become to have the channel of independent error.

Interleaver can have a lot of types, and two kinds on preiodic type and pseudorandom type wherein arranged.Interleaver can also adopt one of two kinds of forms: block structure and convolution structure.The latter is applicable to the convolutional encoding pattern.

The details that in application text, provide the present invention to utilize intersection, but in standard textbook, can find the more particulars of intersecting and going to intersect.According to embodiments of the invention, can implement any intersection and go cross-mode.

The II.Go-CDMA detecting pattern

The Go-CDMA decoding that strengthens

Accurate suboptimum detection algorithm (1.8)-(1.10) of majority logic coding can advantageously be applied to the Go-CDMA decoding block.This be applied under the following condition may be particularly advantageous: low noise environment, Go-CDMA decoding block are decoder stage in the end, perhaps are convolution decoder at it after, and Turbo decoding or piece are decoded.

The enhancing information flow of decoding to subordinate

According to embodiments of the invention, the Go-CDMA decoding block has output z _i, i=1 wherein ... n, and at i=1 ... under the situation of n+1, the z that provides by (1.9) then _iTransmission extra order z _I+1To next decoder stage.Following n=3, the situation of n=5 has special nature:

The situation of n=3.If channel number n equals 3, then the Go-CDMA code signal is:

$s\left(t\right)=\frac{1}{2}\underset{i=1}{\overset{3}{\mathrm{\&Sigma;}}}{d}_i{X}_i\left(t\right)-\frac{1}{2}\left(d_1{d}_2{d}_3{X}_1\left(t\right)X_2\left(t\right)X_3\left(t\right)\right)---\left(1.24\right)$

Consequently, under muting situation, information bit satisfies d _i=z _i, i=1 wherein, 2,3, and d ₁d ₂d ₃=-z ₄Really, for 3 * l Go-CDMA sign indicating number, X _i() is orthonormal, and perpendicular to product X ₁() X ₂() X ₃() therefore, can extract d by decorrelation _i=z _i, i=1 wherein, 2,3, and d ₁d ₂d ₃=-z ₄

The situation of n=5.If channel number n equals 5, then the Go-CDMA code signal is:

$s\left(t\right)=\frac{3}{8}\underset{i=1}{\overset{5}{\mathrm{\&Sigma;}}}{d}_i{X}_i\left(t\right)+\frac{1}{8}\underset{i=1}{\overset{5}{\mathrm{\&Sigma;}}}\underset{j=i+1}{\overset{5}{\mathrm{\&Sigma;}}}\underset{k=j+1}{\overset{5}{\mathrm{\&Sigma;}}}{d}_i{d}_j{d}_k{X}_i\left(t\right)X_j\left(t\right)X_k\left(t\right)---\left(1.25\right)$

$+\frac{3}{8}\left(d_1{d}_2{d}_3{d}_4{d}_5{X}_1\left(t\right)X_2\left(t\right)X_3\left(t\right)X_4\left(t\right)X_5\left(t\right)\right)$

For 3 * l Go-CDMA sign indicating number, X _i() is orthonormal, and X _i(), X _i() X _i() X _k() and x ₁() X ₂() X ₃() X ₄() X ₅Orthogonality is also arranged between ().According to formula (1.25) as can be seen, can information extraction position d by decorrelation _iWith product d _id _jd _kAnd d ₁d ₂d ₃d ₄d ₅

Under the situation of n＞5, point out that following this point is useful, in decorrelation, extract The value of all multinomial coefficients is relatively little.This is because ρ _iThe factor reduces when n＞5 apace, wherein

An example of this situation is n=7.In the case, the Go-CDMA code signal is:

$s\left(t\right)=\frac{5}{16}\underset{i=1}{\overset{7}{\mathrm{\&Sigma;}}}{d}_i{X}_i\left(t\right)-\frac{1}{16}\underset{i=1}{\overset{7}{\mathrm{\&Sigma;}}}\underset{j=i+1}{\overset{7}{\mathrm{\&Sigma;}}}\underset{k=j+1}{\overset{7}{\mathrm{\&Sigma;}}}{d}_i{d}_j{d}_k{X}_i\left(t\right)X_j\left(t\right)X_k\left(t\right)$

$+\frac{1}{16}\underset{i=1}{\overset{7}{\mathrm{\&Sigma;}}}\underset{j=i+1}{\overset{7}{\mathrm{\&Sigma;}}}\underset{k=j+1}{\overset{7}{\mathrm{\&Sigma;}}}\underset{l=k+1}{\overset{7}{\mathrm{\&Sigma;}}}\underset{m=l+1}{\overset{7}{\mathrm{\&Sigma;}}}{d}_i{d}_j{d}_k{d}_l{d}_m{X}_i\left(t\right)X_i\left(t\right)X_k\left(t\right)X_l\left(t\right)X_m\left(t\right)---\left(1.26\right)$

$-\frac{5}{16}{d}_1{d}_2{d}_3{d}_4{d}_5{d}_6{d}_7{X}_1\left(t\right)X_2\left(t\right)X_3\left(t\right)X_4\left(t\right)X_5\left(t\right)X_6\left(t\right)X_7\left(t\right)$

Herein, when i ∈ 2, during n-1}, ρ _iValue is very little.In addition, relate to the required orthogonality of function and be false, except approximate situation, it is coarse therefore making any extraction.

Certainly, suboptimum detecting pattern (1.8)-(1.10) can be used for the situation of other n.According to coefficient ρ among Figure 14 _iRelation between value and the n value, wherein n is an odd number, we can observe under unusual opposite extreme situations, when n=65, from (1.5) first nonzero term and last term still can extract signal amplitude 10% (corresponding to signal power 1%).Under actual conditions, we only need to utilize n≤9 of suboptimum detecting pattern in (1.8)-(1.10), wherein can extract about 10% signal power at the most.

III. the multilevel coding that has the Go-CDMA code level

As above described in the joint, the Go-CDMA decoding not only can provide Go-CDMA information encoded position d ₁, d ₂... d _n∈+1, the estimated value of-1}, but also parity check bit d is provided _N+1:=d ₁d ₂... d _n∈+1, the estimated value of-1}.These soft estimated value is masked as Can be used in any subsequent decoding.The key of this point is to consider that the Go-CDMA encoder is driven by these n+1 positions, and estimates these positions in the Go-CDMA decoding.Any previous coding produces the n+1 position to obtain n information bit as by the model of augmentation with the generation additional parity bit so lump together.When the estimated value of these n+1 positions of decoding, decoding is the encoder that utilizes augmentation.This provides in following coding example, the various Go-CDMA coding techniquess that wherein utilize the Go-CDMA coding to follow.

A) secondary convolution/Go-CDMA coding

Consider to have the standard or the optimization of M shift register The ratio convolution coder.This encoder is discrete time finite memory dynamical system or finite state machine, and m bipolar input arranged, and is masked as u _kM bipolar state is masked as x _kWith n bipolar output, be masked as d _kHerein, k=0,1,2 ... be the discrete time index.Suitable encoder act on set 0, unipolar signal among the+1}.In order to simplify, the special circumstances of the single input of we concentrated discussion m=1, therefore, input is a bipolar data stream, and n output is the n vector current of bipolar coding data, its element is (d _K1, d _K2..., d _Kn) ∈+1 ,-1}.In the case, our formula that provides encoder is

x _k＝Ax _k-1+Bu _k； x ₀＝0，

$d_{\mathrm{kj}}{|}_{j\&Element;\{\mathrm{1,2,3}\&CenterDot;\&CenterDot;\&CenterDot;n\}}=\underset{i=1}{\overset{n}{\mathrm{\&Pi;}}}{g}_i^T{x}_k---\left(1.27\right)$

Herein,

$x_k=\left[\begin{array}{c}{u}_k\\ u_{k-1}\\ u_{k-3}\\ \underset{\&CenterDot;}{\overset{\&CenterDot;}{\&CenterDot;}}\\ u_{k-M}\end{array}\right];A=\left[\begin{array}{ccccc}0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& 0& 0\\ 1& 0& \&CenterDot;\&CenterDot;\&CenterDot;& 0& 0\\ 0& 1& \&CenterDot;\&CenterDot;\&CenterDot;& :& :\\ :& :& \&CenterDot;\&CenterDot;\&CenterDot;& 0& 0\\ 0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& 1& 0\end{array}\right];B=\left[\begin{array}{c}1\\ 0\\ 0\\ :\\ 0\end{array}\right];G^T=\left[\begin{array}{cccc}{g}_1& {\mathrm{<figure-callout\; id="2"\; label="g"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">g</figure-callout>}}_2& \&CenterDot;\&CenterDot;\&CenterDot;& g_n\end{array}\right]--\left(1.28\right)$

And g _i| _{I ∈ 1,2,3...n}}It is the row n vector that bipolar element is arranged.(if u _kWith the element of G be one pole, x then _kAnd d _kAlso be one pole, provide now $d_{\mathrm{kj}}{|}_{j\&Element;\{\mathrm{1,2,3},...n\}}=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\left[g_i^T{x}_k\right]}_2,$ Wherein [] ₂Point out that any addition all is modulus 2 additions.

Now, we this encoder augmentation to output (d is arranged _K1, d _K2..., d _{K (n+1)}) ∈+1 ,-1}'s Rate encoder, wherein d _{K (n+1)}=d _K1d _K2... d _KnMatrix G in the formula (1.28) ^TOnly by additional column g _N+1Augmentation, wherein each element g _{N+1, i}Be " parity check " of above element, g _{N+1, i}=g _{1, i}g _{2, i}... g _{N, i}

Figure 15 provides simple case of this situation of explanation, and wherein solid line is pointed out two shift registers Ratio convolution coder, dotted line point out that augmentation becomes

Convolution coder, therefore, m=1, M=2, n=3.Shown in Figure 15

Convolution coder, but do not have convolution carry-out bit 502x, { 0, the single information bit stream of+1} produces the binary bit stream 502 of 3 vector convolutional encodings to the coding binary form, is masked as d _K1, d _K2And d _K3Then, utilize the d shown in the 502x _K4=d _K1 d _K2 d _K3Carry out augmentation, form

The ratio convolution coder, wherein is modulus 2 add operations.Then, bit stream d _K1, d _K2And d _K3(ignore d _K4) at { d ₁, d ₂, d ₃Intersect (process 600) in the symbolic blocks.Then, by means of 0 →+1 ,+1 →-the 1} conversion, make these convert bipolar binary form to.Then, utilize Go-CDMA cataloged procedure 210, encoded in these bipolar coding positions, produce some extension bits according to the required channel expansion factor.Then, use scrambling at random and shelter sign indicating number 365.

In receiver, after sheltering yard descrambling 368 at random and going to intersect before 610, can in 502x, recover d ₄Estimated value, be masked as

This is the contribution of Go-CDMA decoding embodiment, and it utilizes the last term in the formula (1.5).In receiver, utilize the soft decision Go-CDMA decoded signal of can decoding.Then, decoded signal through conversion+1 → 0 ,-1 →+1} converts binary form to.The data of going to intersect can be fed into carries out soft decision Viterbi convolution decoder

Convolution coder.

Figure 15 only represents the simple case of the embodiment of the invention.Other embodiment can have different m and n value, has the delay shift register 504 of different number M and the different tapped delay of each convolution process 503 to be connected.

Figure 16 A represents that random code in transceiver system 300 and 310 shelters 365 and go to shelter the simple block diagram of general structure before 368, comprising: convolution coder 500 and decoder 510, Go-CDMA encoder 210 and decoder 220, interleaver 600 and remove interleaver 610.Figure 16 B-16D represents the various variations of this embodiment general structure.

Utilize any block encoding after convolutional encoding, for example, Go-CDMA coding is to allow ' RAKE symbol detection ' receiver structure to be used to detect the convolutional encoding sequence and the signature expansion is needed.It is to utilize pure convolutional encoding to make Achilles heel of channel expansion in spread spectrum system that pure convolutional encoding can not be signed detection to bits of coded.In utilizing the Go-CDMA coding of augmentation convolution process, allow to utilize the advantage of ' RAKE symbol detection ' receiver structure to be, former convolutional encoding ratio from

Bring up to There is corresponding performance to improve.Utilize the Go-CDMA coding to have the actual loss of some signal energies, compensate but this loss partly is utilized the parity check bit estimated value that strengthens the Go-CDMA decoding.

Even do not utilize the parity check bit estimated value with the situation that obtains the gain of better convolutional encoding under, the embodiment of the invention is utilized convolutional encoding and combining that Go-CDMA encodes, allow to utilize ' RAKE symbol detection ' receiver structure, otherwise this is impossible, and utilizes the Go-CDMA coding can obtain more healthy and stronger antinoise and the characteristics of signals of better peak value with the ratio (PAP) of average power.

Basic embodiment

In such as 3G and the application of 3G+ employing wireless, the basic embodiment of the embodiment of the invention is to utilize the system of convolution coder in the transmitter, and its parameter is in following scope, m ∈ 1,2,3,4,5,6,7, and 8}, n≤9, M ∈ 3,4,5,6,7,8,9 ... M _MAX, M wherein _MAXIt is the shift register maximum number that current technology can be supported.About the Go-CDMA encoder, its parameter n is in scope n≤9, and l is arranged to the required spreading factor of spread spectrum system, and for example, data rate is 4 chips of 1Mbps in contemporary 3G standard.Preferred n value be odd number value 3,5,7,9}.Interleaver should have sufficient length to handle the sudden mistake of expecting in the transmission channel.In receiver, the compatible interleaver that goes is used in after the soft decision Go-CDMA decoding.Utilizing the convolution interleaver is preferably to select among the convolutional encoding embodiment.Go the output of interleaver to be handled by soft decision Viterbi convolution decoder process, wherein the depth of decode of Li Yonging is to postpone 6 times of tap length in the convolution coder at least.

At given g _N+1The constraint and

Under the ratio sign indicating number, can utilize other convolution code multinomial generator matrix G.The performance metric of these code generators comprises: no code distance, and optimum distance distributes, and best minimum range and optimal spectrum distribute.In a preferred basic embodiment, utilize the G generator matrix of optimizing.Yet,, also can utilize " suboptimum " convolution code multinomial generator matrix G according to embodiments of the invention.

Complicated embodiment

Except the basic embodiment of above convolutional encoding embodiment, below provide complicated or preferred embodiment.

Complicated embodiment such as 3G and 3G+ wireless application is the system that utilizes convolution coder in transmitter, and its parameter is in following scope, m ∈ 1,2,3,4,5,6,7, and 8}, n≤9, M ∈ 3,4,5,6,7,8,9 ... M _MAX, M wherein _MAXIt is the shift register maximum number that current technology can be supported.About the Go-CDMA encoder, parameter n is in scope n≤9, and l is arranged to provide the required spreading factor of spread spectrum system, and for example, data rate is 4 chips of 1Mbps under contemporary 3G standard.Preferred n value be odd number value 3,5,7,9}.Interleaver should have sufficient length to handle the sudden mistake of expecting in the transmission channel.In receiver, the compatible interleaver that goes is used for after the soft decision Go-CDMA decoding.Utilizing the convolution interleaver in this convolutional encoding base embodiment is preferred selection.Go the output of interleaver to be handled by soft decision Viterbi convolution decoder process, the depth of decode of utilization is to postpone 6 times of tap length in the convolution coder at least.

In complicated embodiment structure linguistic context scope, at given g _N+1Constraint and ratio

Under the convolution code, can utilize other convolution code multinomial generator matrix G.In preferred complicated embodiment, utilize the G generator matrix of this optimization.Yet,, also can utilize " suboptimum " convolution code multinomial generator matrix G according to embodiments of the invention.

Figure 17 represents to can be used for to finish the block diagram of the more complicated embodiment of channel expansion.With reference to Figure 17, coding ratio is The output of one or more standard convolutional encoding pieces 500 be fed into two or more Go-CDMA encoding blocks 210 through interleaver 600.The Go-CDMA encoding block is finished the Go-CDMA coding according to any Go-CDMA coding embodiment, and the output of Go-CDMA encoder is connected in series in chronological order, forms the sequence of vectors s (the gray shade district 1000 among Figure 17) of emission, wherein

s＝[s ₁?s ₂ ...s _F] (1.29)

Wherein F represents the Go-CDMA encoder sum that uses.Before transmission, this sequence of vectors of random code concealed blocks 365 scramblings.Convolutional encoding piece 500 can have nothing in common with each other with Go-CDMA encoder 210 or be identical.

In receiver, after the random code in descrambler 368 is sheltered descrambling, the sequence of vectors r of reception (the gray shade district 1100 among Figure 17),

R=s+ interference=[r ₁r ₂... r _F(1.30)

Be divided into sub-vector { r ₁r ₂... r _F.Then, these sub-vectors are fed into Go-CDMA decoder 220 respectively.The output of these Go-CDMA decoders 220 is gone to intersect in piece 610.Then, go the output of interleaver 610 to be fed into separately standard convolutional decoder, wherein the Viterbi decoding algorithm is used for ratio

The augmentation convolution coder.The rearmounted decoder stage of channel in 330 is handled the output of these augmentation convolutional decoders again.

In more complicated embodiment, the spreading factor of communication system is to be provided by following formula

Length (s wherein _i) be vector s _iThe tolerance of chip lengths, and the ratio of β=utilization

The sum of convolution coder.

Shall also be noted that the identical change of interleaver and basic enforcement figure shown in location following Figure 16 A to Figure 16 D that removes the interleaver piece.In addition, when encoder 500 is to have nothing in common with each other and Go-CDMA encoder 210 is when having nothing in common with each other, corresponding difference also appears in the receiver.

In order to further specify this complicated embodiment, this embodiment is divided into two subclasses again, that is, and non-overlapped complexity (NOC) embodiment and overlapping complexity (OC) embodiment.

Non-overlapped complexity (NOC) embodiment

Non-overlapped complexity (NOC) embodiment comprises: augmentation one or more standard convolutional encoding piece and a plurality of Go-CDMA encoding block.

Figure 18 provides the simple case of NOC embodiment, and wherein solid line is pointed out the ratio of four shift registers

Convolution coder, dotted line point out that augmentation arrives Convolution coder, therefore, m=2, M=4, n=6.Shown in Figure 180

Convolution coder, but do not have convolution carry-out bit 502x, the coding binary form 0, the input binary information position u of+1} _k ¹And u _k ², produce 6 bit vector convolutional encoding two-dimensional currents, be masked as d _K1, d _K2, d _K3, d _K4, d _K5And d _K6Utilize the d shown in the 502x _K6+1=d _K1 d _K2 d _K3And d _K6+2=d _K4 d _K5 d _K6, can become these augmentation

The ratio convolution coder.Then, bit stream d _K1, d _K2, d _K3, d _K4, d _K5And d _K6(ignore d _K6+1And d _K6+2) at symbolic blocks { d _K1, d _K2, d _K3And { d _K4, d _K5And d _K6Middle intersected (process 600).These bit streams through conversion 0 →+1 ,+1 →-1} converts bipolar binary form to.Utilize 210 pairs of bipolar coding positions of Go-CDMA cataloged procedure to encode again, produce several extension bits according to the required channel expansion factor.Add then scrambling at random shelter the sign indicating number 365.

In receiver, after sheltering yard descrambling 368 at random and before going to intersect, can in 502x, recover d _K6+1And d _K6+2Estimated value, be masked as

With This is the contribution that utilizes the enhancing Go-CDMA decoding of last term in (1.5).In receiver, utilize soft decision and subsequently go the intersection process, the Go-CDMA decoded signal is decoded.The data of going to intersect are fed into Carry out soft decision Viterbi convolution decoder in the convolution coder.

Figure 18 represents the simple declaration example of this embodiment embodiment.Other embodiment can have different m and n value, and the delay shift register 504 of different number M is arranged, and the different tapped delay of each convolution process 503 connect, by the convolutional encoding position d of each Go-CDMA fgs encoder _KiVarious combination, i=1 wherein, 2 ... n, the convolution coder 500 of different numbers and decoder 510, and the Go-CDMA encoder 210 of different numbers and Go-CDMA decoder 220.A notable feature of NOC embodiment is not have overlapping in utilizing the tapped delay connection that strengthens Go-CDMA decode procedure recovery augmentation position.In other words, in the NOC embodiment, each convolutional encoding position d _KiOnly encoded by a Go-CDMA encoder, i=1 wherein, 2 ... n.

Overlapping complexity (OC) embodiment

In overlapping complexity (OC) embodiment, still can utilize Figure 17 and same general structure shown in Figure 180, but in the OC embodiment, each convolutional encoding position d _KiCan be encoded by a plurality of Go-CDMA encoders, i=1 wherein, 2 ... n.

Provide the simple case of OC embodiment among Figure 19, wherein solid line is pointed out the ratio of four shift registers Convolution coder, and dotted line points out that augmentation arrives

Convolution coder, therefore, m=2, M=4, n=6.Shown in Figure 19 Convolution coder, but do not have convolution carry-out bit 502x, the coding binary form 0, the input binary information position u of+1} _k ¹And u _k ², produce 6 bit vector convolutional encoding two-dimensional currents, be masked as d _K1, d _K2, d _K3, d _K4, d _K5And d _K6Utilize the d shown in the 502x _K6+1=d _K1 d _K2 d _K3 d _K4 d _K5And d _K6+2=d _K2 d _K3 d _K4 d _K5 d _K6, can become these augmentation

The ratio convolution coder.Then, bit stream d _K1, d _K2, d _K3, d _K4, d _K5And d _K6(ignore d _K6+1And d _K6+2) at symbolic blocks { d _K1, d _K2, d _K3, d _K4, d _K5And { d _K2, d _K3, d _K4, d _K5And d _K6Middle intersected (process 600).These bit streams through conversion 0 →+1 ,+1 →-1} converts bipolar binary form to.Utilize 210 pairs of bipolar coding positions of Go-CDMA cataloged procedure to encode again, produce several extension bits according to the required channel expansion factor.Add then scrambling at random shelter the sign indicating number 365.

In receiver, after sheltering yard descrambling 368 at random and before going to intersect, can in 502x, recover d _K6+1And d _K6+2Estimated value, be masked as With

This is the contribution that utilizes the enhancing Go-CDMA decoding of last term in (1.5).In receiver, utilize soft decision and subsequently go the intersection process, the Go-CDMA decoded signal is decoded.The data of going to intersect are fed into Convolution coder carries out soft decision Viterbi convolution decoder.

Figure 19 represents the simple case of this embodiment embodiment.Other embodiment can have different m and n value, the delay shift register 504 of different number M, and the different tapped delay of each convolution process 503 connect, by the overlapping convolutional encoding position d of each Go-CDMA encoder encodes _KiVarious combination, i=1 wherein, 2 ... n, the convolution coder 500 of different numbers and decoder 510, and the Go-CDMA encoder 210 of different numbers and decoder 220.A notable feature of OC embodiment is to have some overlapping in utilizing the tapped delay connection that strengthens Go-CDMA decode procedure recovery augmentation position.This means, in the OC embodiment, each convolutional encoding position d _KiCan be encoded by a plurality of Go-CDMA encoders, i=1 wherein, 2 ... n.

NOC allows greater flexibility with the OC embodiment in the different secondary convolution/Go-CDMA coding mode of configuration, in order that at spreading factor, the given constraint of shift register and data rate realizes best performance down.

In complicated embodiment linguistic context, at given g _N+1Following and the ratio of constraint Under the convolutional encoding, there is the scope (with no code distance, optimum distance distributes, and best minimum range and optimal spectrum distribute and represent) of desirable convolution code multinomial generator matrix G.In preferred complicated embodiment, utilize the G generator matrix of optimizing.Yet, can utilize other " suboptimum " convolution code multinomial generator matrix G.

Other variation

Other variations of present embodiment can comprise:

1) repeated encoding that utilizes low data rate to support

When information data rate was lower than the maximum information data rate of support, repeated encoding perhaps was desirable.Or the information level before being input to convolution coder, or, under the speed that is higher than desired data speed, sample and to realize the repetition of information bit in the coded data level that convolution coder is exported.Perhaps, the output at the Go-CDMA coding can utilize the repetition of extended code chip level.Choosing of number of repetition to reach the compatible extensions factor under the low information data rate.

The example of this situation is an information data rate of supporting 512kbps under 3G WCDMA communication system linguistic context.Required in the case spreading factor is 8 chips.The natural embodiment of present embodiment is to utilize ratio

Convolution coder, wherein m ∈ 1,2,3,4,5,6,7, follow by Go-CDMA coding in 8} and n≤9, and wherein l ∈ 8,16,32,64,128,256}.A kind of may the variation of nature embodiment is to utilize ratio Convolution coder, wherein m ∈ { 1, follow by the Go-CDMA coding, and wherein l=4 and repeated encoding rate are 2 by 2} and n=3.

In receiver, finish identical but opposite process.In addition, to all decoded signal energy additions of duplicate message position, so that in receiver, obtain the estimated value of emission information bit.

2) rate-compatible punctured convolutional (RCPC) the coding replacement standard convolutional encoding of utilizing low data rate to support

It is the standard convolutional encoding of utilizing in rate-compatible punctured convolutional (RCPC) the coding replacement present embodiment that the another kind of this embodiment changes, in order that support the different information data rates of identical convolution coded hardware structure.Utilize above-mentioned repeated encoding also can replenish this variation.

In receiver, finish identical but opposite process (RCPC decoding) to received signal.

B) secondary Turbo/Go-CDMA coding

Consider standard or optimization respectively Ratio Turbo encoder and ratio are Two parallel recurrence system convolution (RSC) encoder, n wherein _TurboIt is the coding carry-out bit number of Turbo encoder.Each rsc encoder has M shift register, and before second rsc encoder coding, the pseudorandom intersection is added to information bit.Fig. 7 represents this situation.The Turbo encoder also can be regarded parallel convolution (PCC) encoder that connects as.Therefore, each rsc encoder in the Turbo encoder is discrete time finite memory dynamical system or finite state machine, and m bipolar output is wherein arranged, and is masked as u _kM bipolar state is masked as x _kWith n bipolar output, be masked as d _k ^ζ, wherein { 1,2} represents rsc encoder 1 and rsc encoder 2 in the Turbo encoder to ζ ∈.Herein, k=0,1 ... be the discrete time index.Fig. 8 represents the basic structure of rsc encoder.Suitable encoder act on set 0, the unipolar signal the among+1}.

Because each rsc encoder in the Turbo encoder can be represented with suitable convolution coder, this is an alternative embodiment of the invention, and wherein secondary convolution/Go-CDMA coding can be added to each rsc encoder in the Turbo encoder.This just allows us to obtain the enhancing of secondary Turbo/Go-CDMA coding.

In order to simplify, the single input condition that our concentrated discussion is special, m=1 wherein, therefore, input is a bipolar data stream, and n the n bit vector stream that output is the bipolar coding data, its element $({\mathrm{<figure-callout\; id="1"\; label="d\; k"\; filenames="A0182193800571.png,A0182193800631.png"\; state="\{\{state\}\}">d</figure-callout>}}_k^1,{\mathrm{<figure-callout\; id="2"\; label="d\; k"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">d</figure-callout>}}_k^2,{\mathrm{<figure-callout\; id="3"\; label="d\; k"\; filenames="A0182193800631.png"\; state="\{\{state\}\}">d</figure-callout>}}_k^3\&CenterDot;\&CenterDot;\&CenterDot;,d_{\mathrm{kn}}^{\mathrm{\&xi;}})\&Element;\{+1,-1\}.$ For the Turbo encoder, we advise that identical sign indicating number is used for this two rsc encoders.In addition, in each rsc encoder, suitable convolution coder is arranged.Based on these hypothesis, do not lose generality, we can provide the formula in (1.27) and (1.28) to rsc encoder, wherein

x _k＝Ax _k-1+Bu _k；x ₀＝0，

$d_{\mathrm{kj}}^{\mathrm{\&xi;}}{|}_{j\&Element;\{\mathrm{1,2,3}\&CenterDot;\&CenterDot;\&CenterDot;n\}}=\underset{i=1}{\overset{n}{\mathrm{\&Pi;}}}{g}_i^T{x}_k---\left(1.32\right)$

Herein,

$x_k=\left[\begin{array}{c}{u}_k\\ u_{k-1}\\ u_{k-3}\\ \underset{\&CenterDot;}{\overset{\&CenterDot;}{\&CenterDot;}}\\ u_{k-M}\end{array}\right];A=\left[\begin{array}{ccccc}0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& 0& 0\\ 1& 0& \&CenterDot;\&CenterDot;\&CenterDot;& 0& 0\\ 0& 1& \&CenterDot;\&CenterDot;\&CenterDot;& :& :\\ :& :& \&CenterDot;\&CenterDot;\&CenterDot;& 0& 0\\ 0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& 1& 0\end{array}\right];B=\left[\begin{array}{c}1\\ 0\\ 0\\ :\\ 0\end{array}\right];G^T=\left[\begin{array}{cccc}{g}_1& {\mathrm{<figure-callout\; id="2"\; label="g"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">g</figure-callout>}}_2& \&CenterDot;\&CenterDot;\&CenterDot;& g_n\end{array}\right]--\left(1.33\right)$

And g _i| _{I ∈ 1,2,3 ... n}}It is the row n vector that bipolar element is arranged.(if u _kWith the element of G be one pole, x then _kAnd d _k ^ζAlso be one pole, provide now $d_{\mathrm{kj}}^{\mathrm{\&xi;}}{|}_{j\&Element;\{\mathrm{1,2,3},...n\}}=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\left[g_i^T{x}_k\right]}_2,$ Wherein [] ₂Point out that any addition all is modulus 2 additions.Therefore, utilize the PCC configuration of convolution coder in (1.32) and (1.33), we can realize ratio The Turbo encoder.Because in each convolution coder, because in its rsc encoder system path is arranged; Therefore, the PCC encoder in the Turbo encoder has two system paths, rather than as has only a system path in standard Turbo coding.This conversion has following advantage:

The Turbo coder structure is transformed into the PCC coder structure of two identical standard convolution coders.So we can use Go-CDMA coding now and be strengthened on separately the convolution coder, are entitled as " secondary convolution/Go-CDMA coding " described mode according to last joint acceptance of the bid.

In addition, in the PCC coder structure, two system paths are arranged, and a system path is only arranged in the Turbo coder structure.

For this purpose, we become to have output to the convolution coder augmentation of (1.32) and (1.33) now $(d_{\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="A0182193800571.png,A0182193800631.png"\; state="\{\{state\}\}">k</figure-callout>}1}^{\mathrm{\&xi;}},{\mathrm{<figure-callout\; id="2"\; label="d\; k"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">d</figure-callout>}}_k^2,{\mathrm{<figure-callout\; id="3"\; label="d\; k"\; filenames="A0182193800631.png"\; state="\{\{state\}\}">d</figure-callout>}}_k^3\&CenterDot;\&CenterDot;\&CenterDot;,d_{k(n+1)}^{\mathrm{\&xi;}})\&Element;\{+1,-1\}$ Rate encoder, wherein $d_{k(n+1)}^{\mathrm{\&xi;}}={\mathrm{<figure-callout\; id="1"\; label="d\; k"\; filenames="A0182193800571.png,A0182193800631.png"\; state="\{\{state\}\}">d</figure-callout>}}_k^1{\mathrm{<figure-callout\; id="2"\; label="d\; k"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">d</figure-callout>}}_k^2\&CenterDot;\&CenterDot;\&CenterDot;d_{\mathrm{kn}}^{\mathrm{\&xi;}}.$ (by utilizing Go-CDMA coding and improved decoding, as in above several joints, describing in detail).(1.33) the matrix G in ^TOnly by additional row g _N+1Augmentation, wherein element g _{N+1, i}Be " odd-even check " of above element, g _{N+1, j}=g _{1, i}g _{2, i}... g _{N, i}

Being intended that of this embodiment of the present invention is applied to each convolution coder in the standard Turbo coding structure to title for " secondary convolution/Go-CDMA coding " goes up all embodiment in the joint.

In receiver, " parity check bit " of from improved Go-CDMA decoding, deriving augmentation.These " parity check bits " can be used for strengthening the decoding algorithm performance in the Turbo decoder.In addition, because secondary Turbo/Go-CDMA coding method allows the transmission of two system paths, rather than a system path in the standard Turbo coding, we obtain further to improve to the overall code capacity of standard Turbo coding mode.

C) second-order block/Go-CDMA coding

Based on analog study and the observation to formula (1.5), we infer is under the situation of even number at n in the Go-CDMA system, when access channel (being input to encoder) number when n channel is increased to n+1 channel, performance does not worsen.Now, utilize the Go-CDMA coding of describing in this supposition and above two joints to strengthen, construct second-order block/Go-CDMA coding.The advantage that obtains of expectability is not big as under convolution/Go-CDMA coding or the Turbo/Go-CDMA coding situation for we, and really, the relative merit of expection reduces with the increase of block length.

Embodiments of the invention are the second-order block/Go-CDMA coding modes under n the information bit situation, and wherein n is an even number.Block code only increases a parity check bit 1, is referred to as first parity check bit, to obtain odd number n+1 position.Now, this odd bits is applied to the Go-CDMA coding mode.Increase a parity check bit and can not cause the performance loss of essence for other Go-CDMA coding, because n is an even number, but the estimated value of parity check bit can improve the estimated value of information bit.

The Go-CDMA that the utilization of Go-CDMA code level strengthens detects, and the additional parity bit of its estimation coding n+1 position is referred to as second parity check bit.N is more little, and then the influence of first second parity check bit is just big more.Value n=2 provides maximum gain, provides than ordinary circumstance more detailed description now.

Embodiments of the invention are, when group of information bits is made into b * n piece (matrix D), wherein b is a positive integer, and n is even positive integer.Relevant with the value of b, can implement various linear block codings to information bit and select, realize first parity check bit relevant with every row among the D, expand row for information bit matrix D, be masked as p.Every row to [Dp] expands second parity check bit again, and it is the product of row element under bipolar element situation, and is modulus 2 additions under one pole element situation, expands an additional column for [Dp], is masked as φ, as B=[Dp φ] shown in.Then, this matrix B of the Go-CDMA of enhancing decoding estimation.The best decode procedure of estimation original information bits is that PRML detects.This is for b, and the situation that n is very little is very simple.

An example is n=2, the situation of b=2.Information bit is processed in 4 pieces.These information bits are labeled as { d ₁, d ₂, d ₃, d ₄, first parity check bit of interpolation is { p ₁, p ₂And second parity check bit be { φ ₁, φ ₂, they appear among the matrix format B, and then information of Shou Jiing and interpolation parity check bit are exactly as follows:

$B=\left[\begin{array}{cccc}{d}_1& d_2& {\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="A0182193800571.png,A0182193800631.png"\; state="\{\{state\}\}">p</figure-callout>}}_1& {\mathrm{\&phi;}}_1\\ d_3& d_4& {\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">p</figure-callout>}}_2& {\mathrm{\&phi;}}_2\end{array}\right]---\left(1.34\right)$

A kind of result who bipolar information bit is implemented the linear block coding is

p ₁＝d ₁d ₃d ₄；p ₂＝d ₁d ₂d ₃； (1.35)

Therefore,

φ ₁＝p ₁d ₂d ₁＝d ₂d ₃d ₄；φ ₂＝p ₂d ₃d ₄＝d ₁d ₂d ₄ (1.36)

In this embodiment, each parity check bit is supported an information specific position d _i, i={1 wherein, 2,3,4}.Under the situation of one pole information bit, d ^T={ d ₁d ₂d ₃d ₄, then

${\&lsqb;d^{T}G\&rsqb;}_2=\&lsqb;d_1{d}_2{d}_3{d}_4{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="A0182193800571.png"\; state="\{\{state\}\}">p</figure-callout>}}_2{\mathrm{\&phi;}}_2{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="A0182193800571.png,A0182193800631.png"\; state="\{\{state\}\}">p</figure-callout>}}_1{\mathrm{\&phi;}}_1\&rsqb;=:d_{\mathrm{aug}}^T---\left(1.37\right)$

Wherein [] ₂Illustrate that once more addition is modulus 2 arithmetical operations, and G is a generator matrix,

$G=\left[\begin{array}{cccccccc}1& 0& 0& 0& 1& 1& 1& 0\\ 0& 1& 0& 0& 1& 1& 0& 1\\ 0& 0& 1& 0& 1& 0& 1& 1\\ 0& 0& 0& 1& 0& 1& 1& 1\end{array}\right]---\left(1.38\right)$

Under the best decoding was selected, above example utilized PRML to detect, and checks the information bit that receives With 2 of D ⁴Individual relation that may figure, and with the relation of B.Other suboptimal solution code calculations of above example comprise the standard utilization of hard decision base decoding, and are as described below.

The corresponding bug list that the decoding of hard decision base utilizes parity matrix and has the mistake figure that can revise.

In above example, corresponding parity check matrix H is

$H=\left[\begin{array}{cccccccc}1& 1& 1& 0& 1& 0& 0& 0\\ 1& 1& 0& 1& 0& 1& 0& 0\\ 1& 0& 1& 1& 0& 0& 1& 0\\ 0& 1& 1& 1& 0& 0& 0& 1\end{array}\right]---\left(1.39\right)$

According to d ^T _AugH ^T, below provide 2 ⁴Individual possible data and relevant d _AugThe bug list that produces.In addition, give can revise might one and the corresponding mistake figure of some two information bit mistake.

Fault mistake figure

$\left[\begin{array}{cccc}0& 0& 0& 0\\ 1& 1& 1& 0\\ 1& 1& 0& 1\\ 1& 0& 1& 1\\ 0& 1& 1& 1\\ 1& 0& 0& 0\\ 0& 1& 0& 0\\ 0& 0& 1& 0\\ 0& 0& 0& 1\\ 0& 0& 1& 1\\ 0& 1& 0& 1\\ 1& 0& 0& 1\\ 0& 1& 1& 0\\ 1& 0& 1& 0\\ 1& 1& 0& 0\\ 1& 1& 1& 1\end{array}\right]$ $\left[\begin{array}{cccccccc}0& 0& 0& 0& 0& 0& 0& 0\\ 1& 0& 0& 0& 0& 0& 0& 0\\ 0& 1& 0& 0& 0& 0& 0& 0\\ 0& 0& 1& 0& 0& 0& 0& 0\\ 0& 0& 0& 1& 0& 0& 0& 0\\ 0& 0& 0& 0& 1& 0& 0& 0\\ 0& 0& 0& 0& 0& 1& 0& 0\\ 0& 0& 0& 0& 0& 0& 1& 0\\ 0& 0& 0& 0& 0& 0& 0& 1\\ 1& 1& 0& 0& 0& 0& 0& 0\\ 1& 0& 1& 0& 0& 0& 0& 0\\ 1& 0& 0& 1& 0& 0& 0& 0\\ 1& 0& 0& 0& 1& 0& 0& 0\\ 1& 0& 0& 0& 0& 1& 0& 0\\ 1& 0& 0& 0& 0& 0& 1& 0\\ 1& 0& 0& 0& 0& 0& 0& 1\end{array}\right]$

(1.40)

Other embodiments

Other possibility embodiments of the present invention are included in the Go-CDMA coded system that strengthens error correction utilizes the whole bag of tricks of parity check bit, but does not change the spreading factor of Go-CDMA sign indicating number.The possible example of these methods is:

Example A: consider to have two information bit d ₁And d ₂, and first parity check bit is p ₁=d ₁And second parity check bit be φ ₁=d ₂Situation.In this case, each parity check bit is supported an information bit.

Example B: considering has 9 channels in 9 * l Go-CDMA system, 8 information bit { d only encode ₁... d ₈And 1 parity check bit p ₁, p wherein ₁=d ₁d ₂d ₃d ₄Therefore, utilize the intrinsic polynomial construction of Go-CDMA code signal, second parity check bit is φ ₁=p ₁d ₁d ₂... d ₈=d ₅d ₆d ₇d ₈In this example, first parity check bit is supported preceding 4 information bits, and second parity check bit is supported back 4 information bits.

Example C: consider in 9 * l Go-CDMA system, to have 9 channels, coding 8 * 8 block message position D _{8 * 8}And the situation of a parity check vector p, wherein

$D_{8\&times;8}=\left[\begin{array}{cccccccc}{d}_1& d_2& d_3& d_4& d_5& d_6& d_7& d_8\\ d_9& d_{10}& d_{11}& d_{12}& d_{13}& d_{14}& d_{15}& d_{16}\\ d_{17}& d_{18}& d_{19}& d_{20}& d_{21}& d_{22}& d_{23}& d_{24}\\ d_{25}& d_{26}& d_{27}& d_{28}& d_{29}& d_{30}& d_{31}& d_{32}\\ d_{33}& d_{34}& d_{35}& d_{36}& d_{37}& d_{38}& d_{39}& d_{40}\\ d_{41}& d_{42}& d_{43}& d_{44}& d_{45}& d_{46}& d_{47}& d_{48}\\ d_{49}& d_{50}& d_{51}& d_{52}& d_{53}& d_{54}& d_{55}& d_{56}\\ d_{57}& d_{58}& d_{59}& d_{60}& d_{61}& d_{62}& d_{63}& d_{64}\end{array}\right],\left(1.41\right)$

p＝p ₁?p ₂?p ₃?p ₄?p ₅?p ₆?p ₇?p ₈]

Each parity check bit inspection message position piece D _{8 * 8}In one row, that is, and p ₁Be that inspection message ranks p ₁, p ₉, p ₁₇, p ₂₅, p ₃₃, p ₄₁, p ₄₉And p ₅₇The parity check bit of last information bit.Then, this block of information of parity check vector augmentation forms 8 * 9 matrix-blocks, and wherein the 9th row are parity check vector p, as [D _{8 * 8}P].

Utilize this method, implement 2 dimension checksum coding patterns, wherein first parity check bit is checked the vertical direction of 8 * 8 information bit matrixs, and the second parity check bit inspection [D _{8 * 8}D _{8 * 8}] horizontal direction of matrix.

D) secondary convolution or piece/many yards Go-CDMA codings

In this embodiment, can finish convolutional encoding or Turbo coding or block encoding according to the above embodiments.

Many yards Go-CDMA (MC-Go-CDMA) coding refers to and utilizes the Go-CDMA sign indicating number as " nonlinear tolerance " Go-CDMA sign indicating number, as being " described in the related U.S. patent application of Methodand Apparatus for Non-Linear Code-Division Multiple AccessTechnology; application on October 5 calendar year 2001, to be incorporated in that this is for reference at title.

Described at above related application, when Go-CDMA sign indicating number during as the extended code in the MC-CDMA structure, many level composite signal of formation is allowed big nonlinear distortion.Therefore, the MC-Go-CDMA decoded signal is more reliable.Utilize the reliability of this increase, provide an alternative embodiment of the invention herein.

The convolution of above embodiment or block encoding are to be used in combination with MC-Go-CDMA coding in the transmitter.In receiver, at the signal r (t) that detects reception afterwards, wherein r (t)=s (t)+interference is according to the estimated value of following computing derivation parity check bit

${\hat{d}}_{n+1}=\underset{0}{\overset{t}{\&Integral;}}\mathrm{sign}\left(r\left(t\right)\right)X_1\left(\mathrm{\&tau;}\right)X_2\left(\mathrm{\&tau;}\right)\&CenterDot;\&CenterDot;\&CenterDot;X_n\left(\mathrm{\&tau;}\right)\mathrm{d\&tau;}$

Or (1.42)

${\hat{d}}_{n+1}=\underset{0}{\overset{t}{\&Integral;}}{\mathrm{sgn}}_{\&PlusMinus;}\left(r\left(t\right)\right)X_1\left(\mathrm{\&tau;}\right)X_2\left(\mathrm{\&tau;}\right)\&CenterDot;\&CenterDot;\&CenterDot;X_n\left(\mathrm{\&tau;}\right)\mathrm{d\&tau;}$

Therefore, in this embodiment of the present invention, the soft decoded bits of formation Be used further to convolution decoder or Turbo decoding or piece decoding among some embodiment herein.

Error correction computing in pre-coded and the rearmounted decoder stage can be made up, assimilation or with the Go-CDMA replacement that has error correction coding in the foregoing description.

The sign indicating number lattice of the uniform enconding pattern of e) introducing in this patent are represented

Linear finite state coding mode can be expressed as the form of trellis code.Therefore, for the sake of completeness, we provide the following sign indicating number lattice of the basic coding pattern that proposes in this patent and represent.

Secondary convolution/Go-CDMA encodes as grid coding

Consider to have the standard or the optimization of M shift register

The ratio convolution coder.This encoder is discrete time finite memory dynamical system or finite state machine, and m bipolar input wherein arranged, and is masked as u _kM bipolar state is masked as x _kWith n bipolar output, be masked as d _kHerein, k=0,1,2 ... be the discrete time index.In order to simplify, our concentrated discussion n output forms the special circumstances of the n vector current of bipolar coding data, and element is (d _K1, d _K2..., d _Kn) ∈+1 ,-1}.

N output is the Go-CDMA coding.Because have 2 ⁿIndividual possible n bit vector is input to the Go-CDMA encoder, just has 2 ⁿIndividual possible output sequence.Therefore, the Go-CDMA coding is that the n bit vector transforms to output sequence.This conversion can be regarded the part of modulated process as.Convolution coder is the effective status encoder, and it constitutes special trellis encoder with this conversion.

A) optimization of trellis code

The optimization of trellis code relates to the combined optimization of convolution code and modulation (conversion).Traditionally, when the design trellis code, take good convolution code, thereby make conversion optimization.Yet in convolution/Go-CDMA coding, conversion is fixing by the Go-CDMA coding.So the optimization of convolution code is depended in the optimization of trellis code.

B) trellis decode

Convolution/Go-CDMA coding is regarded as grid coding, and then Go-CDMA decoding and convolution decoder can replace with the single step trellis decode.

The sequence transmission that receives in the signaling interval is by one group 2 ⁿIndividual filter.Each filter and the output sequence coupling that one of them is possible.The output of filter is used for branch's matrix of computing grid.Decoding utilizes the Viterbi algorithm to finish.

Though trellis decode is possible, yet the convolution decoder after the Go-CDMA decoding remains preferred in practice.In the Go-CDMA decoding, only need n+1 filter, and in trellis decode, need 2 ⁿIndividual filter.

Secondary Turho/Go-CDMA coding is as grid coding

In Turbo/Go-CDMA coding, utilize with the convolutional encoding situation under identical basic code lattice represent, but under the linguistic context of convolution (PCC) encoder of parallel connection.

Second-order block/Go-CDMA encodes as grid coding

In piece/Go-CDMA coding mode, can consider the binary of generator matrix G and parity-check matrix H

Linear block code generated code lattice are represented.During each encoded interval, the message of m information bit is displaced in the encoder memory and is encoded into the code word of n sign indicating number position.Form n sign indicating number position, and at n bit time internal shift to channel.So coding span Γ is limited, it comprises n+1 time instance, wherein

Γ＝{1，2，...，N} (1.43)

Therefore, utilize time span Γ to go up n section lattice figure, can represent binary

Linear block code.In addition, because the n of block code encoder output is the Go-CDMA coding, and owing to have 2 ⁿIndividual possible n bit vector is input to the Go-CDMA encoder, so, just have 2 ⁿIndividual possible output sequence.Therefore, the Go-CDMA coding is that the n bit vector transforms to output sequence.This conversion can be regarded the part of modulated process as.The linear block encoder is the finite state encoder that trellis code is arranged, and combines with this Go-CDMA conversion to constitute special trellis encoder.

A) optimization of trellis code

The optimization of trellis code requires the combined optimization of linear block code and modulation (conversion).Traditionally, when the design trellis code, take the favorable linearity block code, thereby make conversion optimization.Yet in piece/Go-CDMA coding, conversion is fixing by the Go-CDMA coding.So the optimization of linear block code is depended in the optimization of trellis code.

B) trellis decode

Piece/Go-CDMA coding is regarded as grid coding, and then Go-CDMA decoding and piece decoding can replace with the single step trellis decode.

The sequence transmission that receives in the signaling interval is by one group 2 ⁿIndividual filter.Each filter and a possible output sequence coupling.The output of filter is used for branch's matrix of computing grid.Decoding utilizes maximum-likelihood algorithm to finish.

Though can implement trellis decode, the piece decoding after the Go-CDMA decoding may remain preferred in practice.In the Go-CDMA decoding, need n+1 filter, and in trellis decode, need 2 ⁿIndividual filter.

The trellis coded modulation (TCM) of the uniform enconding pattern of f) introducing in this patent

Utilization in the trellis coded modulation technology that nineteen eighty-two proposes, can also be modulated Go-CDMA coding output in the embodiment of the invention based on Ungerboeck.Utilize the TCM in the transmitter 3 to 6dB coding gain is provided for each embodiment of the present invention.Usually, only in receiver, utilize and just obtain this coding gain when detecting the corresponding TCM detecting pattern of (MLD) based on PRML.The MLD pattern has the complexity of enforcement, and complexity is to increase by exponential manner with the signal mode number that may detect.Therefore, possible signal mode number and the current enforcement of hardware technology physical constraints TCM in the mini system of limited channel number in the output.

According to another embodiment of the present invention, conventional TCM pattern can be used for conversion strengthens coding mode according to the present invention output.

Utilize the TCM conversion of non-linear block codewords matrix

One embodiment of the present of invention be utilize the TCM technology from the transmitter output transform that strengthens coding mode according to the present invention to non-linear code word matrix.Non-linear code word matrix can constitute by means of following method:

The Levenshtein building method;

| u|u v| building method; And

| u|u v| building method, wherein u and v are the non-overlapped parts of Go-CDMA coding output.

The receiver of this embodiment utilizes the PRML detection technique with bits of coded and additional parity check bit that estimation receives, is used in augmentation error-decoded process of the present invention then.Perhaps, the MLD process in the receiver can expand to before the error correction coding PRML of transmitted data signal estimated value is provided.

Actual coding embodiment

In Figure 17, be expressed as three grades of possible actual embodiments according to channel expansion of the present invention and next code embodiment.In this joint, possible actual embodiment is described.

Figure 20 represents the communication system functionality block diagram of actual coding embodiment of the present invention, comprises channel error correcting encoder (330a) and decoding (330b), channel expansion (210) and go expansion (220) and modulate forward transform (211) and reciprocal transformation (221).Between antenna and channel error correcting encoder and decoding block 330, comprise piece 330, each level course of system can be considered to different code level as shown in the figure.

An actual embodiment is a communication system shown in Figure 20 in the embodiment of the invention, wherein channel expansion (210) and go the expansion (220) functional block comprise two code level.First code level (210a) and decoder stage (220a) mainly contain the function that the error correction coding is provided, and second code level (210b) and decoder stage (220b) have first order error correction coding (210a and 220a) function and channel expansion function in the further reinforcement communication system.As shown in figure 20, the output of second level coding 210b can randomly be done further conversion by means of the trellis coded modulation technology before the Channel Modulation in the channel expansion functional block in 320.Opposite process also randomly is applicable to receiving terminal.

The another kind of actual embodiment of the embodiment of the invention is the communication system that is similar to system shown in Figure 20, wherein channel error correcting encoder (330a) and decoding (330b) functional block make up with first order encoding block (210a) and decoding block (220a) respectively, form stronger error correction encoding block.In this actual embodiment, modulation forward transform (211) and reciprocal transformation (221) are also chosen wantonly in system.

For two actual embodiments of the present invention, definite decorrelated signals pattern is arranged from the output of second level coding 210b.In receiver, utilize mathematical method described herein, can this decorrelated signals pattern of decorrelation, the complexity of this mathematical method increases according to linear mode with the data channel number.

Though we are each specific embodiment of the present invention openly, this area general professional can understand, under the condition that does not depart from spirit and scope of the invention, can do some variations to these embodiment.For example, though described various functions and code level, should be understood that one or more levels can be combined into the individual feature piece, on the contrary, the individual feature piece can resolve into one or more levels.Any embodiment of above-mentioned functions and technology is within the scope of the invention, implements to have nothing to do in any level with this function or technology, and be that enforcement is irrelevant in how many levels also with this function or technology.Be to be further appreciated that, utilize the mathematical operation of Go-CDMA sign indicating number with error correction coding and matrix and the matrix operation can be with hardware or software implementation.Under the situation of software implementation, software instruction and data can be included in the useable medium of computer, and store memory in the communicator into.Software instruction can comprise the control logic that processor or other hardware are carried out, and makes communicator based on Go-CDMA sign indicating number Code And Decode data-message, and wherein the error correction coding is described in the accompanying drawing.When utilizing hardware or firmware to implement, utilize the Go-CDMA sign indicating number with error correction coding and the mathematical operation and the matrix operation of matrix to provide by the logic on one or more chips, or can be burned in the EEPROM as program command and data.Be to be further appreciated that, Go-CDMA sign indicating number and matrix with error correction can be retrieved and utilize by system, can store memory into, are included in the hardware, receive from external source, or derive the storage data inside and outside system or from hardware, produce such as other hardware or memory.

Claims (38)

1. A method for encoding and transmitting data signals, comprising: In a first coding stage, an error correction coding mode is used to generate coded bits based on the data signal; In a second stage, re-encoding the coded bits based on a non-linear block error correction coding scheme to achieve an output with a detectable pattern; and A channel extension signal based on the output is transmitted. 2. A method according to claim 1, wherein the first and second level encoding modes are selected to enhance error correction during decoding. 3. The method according to claim 1, wherein the second stage of non-linear block error correction coding involves a Go-CDMA matrix multiplication operation. 4. The method according to claim 1, wherein the first level error correction coding mode is a convolutional coding mode. 5. The method according to claim 1, wherein the first level error correction coding mode is a Turbo coding mode. 6. The method according to claim 1, wherein the first level error correction coding mode is a block coding mode. 7. The method of claim 1, further comprising: interleaving encoding bits prior to the second stage of encoding. 8. The method of claim 1, further comprising: interleaving the outputs prior to transmitting. 9. The method according to claim 1, wherein the first stage comprises at least one encoder. 10. A method according to claim 1, wherein the second stage comprises at least one non-linear block error encoder. 11. The method according to claim 1, wherein the data coding of the channel error correction target is included in the first stage of channel spreading coding. 12. The method according to claim 1, wherein the first and second coding levels are represented by trellis coding patterns. 13. The method of claim 1, wherein the output is modulated using a trellis code modulation technique. 14. A method according to claim 1, wherein the output is transformed to a non-linear codeword matrix of a larger code distance using a trellis code modulation technique. 15. A method of decoding a channel extension encoded signal comprising: Receive channel extension signal; In a first decoding stage, based on a non-linear error correction coding scheme, decorrelating the received signal to recover coded bit estimates in blocks of n bits, and estimating additional parity bits from the received signal; and In a second decoding stage, based on the error correction decoding mode, the encoded n bits and the appended parity bits are decoded to recover the data signal. 16. A method according to claim 15, wherein the channel extension signal is received and decorrelated in a multi-receiver structure called a RAKE receiver. 17. A method according to claim 16, wherein the estimated values of the data bits and parity bits from each RAKE receiver structure are soft values. 18. The method according to claim 16, wherein the first stage nonlinear block error correction decoding involves a Go-CDMA matrix decorrelation operation. 19. The method according to claim 15, wherein the error correction decoding mode is a convolutional code decoding mode. 20. The method according to claim 15, wherein the error correction decoding mode is a Turbo code decoding mode. 21. The method according to claim 15, wherein the error correction decoding mode is a block code decoding mode. 22. The method of claim 15, further comprising deinterleaving the received channel spread signal bits prior to the first decoding stage. 23. The method of claim 15, further comprising: deinterleaving the bits after the first decoding stage. 24. The method according to claim 15, wherein the channel extension signal is decoded in the first stage of decoding using at least one non-linear block error decoder. 25. The method according to claim 15, wherein the second error correction decoding stage comprises at least one decoder. 26. A method according to claim 15, wherein any decoding associated with channel error coding targets is integrated into the first decoding stage. 27. The method according to claim 16, wherein the first stage nonlinear block error correction decoding involves Go-CDMA matrix decorrelation operations. 28. The method of claim 16, wherein the second stage decoder is a convolutional code decoder. 29. The method according to claim 16, wherein the second stage decoder is a turbo code decoder. 30. The method of claim 16, wherein the second stage decoder is a block code decoder. 31. The method of claim 16, further comprising deinterleaving the received channel spread signal bits prior to the first decoding stage. 32. The method of claim 16, further comprising: deinterleaving the bits after the first decoding stage. 33. The method according to claim 16, wherein the channel extended signal is decoded in the first stage decoder using at least one non-linear block error decoder. 34. The method of claim 16, wherein the second decoding stage includes at least one decoder. 35. A method according to claim 16, wherein any decoding associated with the channel error correction coding object is integrated into the first decoding stage. 36. The method according to claim 15, wherein the decoding of the channel spread coded signal is based on a maximum likelihood detection algorithm. 37. The method according to claim 16, wherein the decoding of the channel spread coded signal is based on a maximum likelihood detection algorithm. 38. The method according to claim 35, wherein the decoding of the channel spread coded signal is based on a maximum likelihood detection algorithm in trellis decoding mode.

Images