[go: up one dir, main page]

HK1111835A - Method and apparatus for computing soft decision input metrics to a turbo decoder - Google Patents

Method and apparatus for computing soft decision input metrics to a turbo decoder Download PDF

Info

Publication number
HK1111835A
HK1111835A HK08102436.0A HK08102436A HK1111835A HK 1111835 A HK1111835 A HK 1111835A HK 08102436 A HK08102436 A HK 08102436A HK 1111835 A HK1111835 A HK 1111835A
Authority
HK
Hong Kong
Prior art keywords
img
symbols
soft decisions
group
gif
Prior art date
Application number
HK08102436.0A
Other languages
Chinese (zh)
Inventor
N．辛德伍沙雅那
Original Assignee
高通股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 高通股份有限公司 filed Critical 高通股份有限公司
Publication of HK1111835A publication Critical patent/HK1111835A/en

Links

Description

Method and apparatus for computing soft decision input metrics for a Turbo decoder
The present application is a divisional application of the chinese patent application entitled "method and apparatus for calculating soft decision input metric for Turbo decoder" of the invention No. 01808893.7, 3/7/2001.
Technical Field
The present invention relates generally to the field of communications, and more particularly to the computation of soft decision input metrics for turbo decoders.
Background
The transmission of digital data is inherently susceptible to interference, which may introduce errors in the transmitted data. Error detection schemes have been proposed to determine as much as possible whether errors have been introduced in the transmitted data. For example, it is common practice to transmit data in packets and to add a Cyclic Redundancy Check (CRC) field, for example of length 16 bits, on each packet, which carries the checksum of the data of the packet. When a receiver receives data, the receiver calculates the same checksum on the received data and checks whether the calculation result is consistent with the checksum in the CRC field.
When the transmission data is not used in real time, it is possible to request retransmission of erroneous data when an error is detected. However, when real-time transmission is being made, such as a voice call on a traditional telephone or cellular telephone, a video conference call, etc., it is possible to request retransmission.
Convolutional coding has been introduced to allow receivers of digital data to correctly determine the transmitted data even when errors have occurred during transmission. The convolutional encoding introduces redundancy in the transmitted data and packs the transmitted data into packets in which the value of each bit depends on the previous bit in the sequence. Thus, when errors occur, the receiver is also able to deduce the original data by tracing back possible sequences in the received data.
To further improve the performance of the transmission channel, some coding schemes include interleaving, which mixes the order of bits in the packets during encoding. Thus, when interference destroys adjacent bits during transmission, the effect of the interference is spread over the entire original packet and can be easily overcome by the decoding process. Other improvements may include encoding multiple components of a packet code in parallel or in series more than once. For example, it is well known in the art to use error correction methods that use at least two convolutional encoders in parallel. Such parallel encoding is generally referred to as turbo encoding.
turbo coding is a serial or parallel concatenation of two or more constituent encoders separated by one or more code interleavers. Turbo codes are often decoded using relatively efficient iterative algorithms to achieve low error rates where the signal-to-noise ratio (SNR) approaches the shannon limit. The interleaver and deinterleaver are inserted between a component encoding encoder and decoder, respectively.
As described above, the interleaver in a turbo encoder spreads the encoded word output from the encoder so that the individual bits of a given encoded word are separated from each other and transmitted at different times. Thus, individual bits of a given code experience independent fading, such that the bits affected by an error burst belong to different code words. At the receiver, the received samples are deinterleaved prior to decoding. The effect of the error burst is thus spread over the message, making it possible to recover the data with the original error correction coding. the performance of turbo coding depends on the length and structure of the code interleaver used. Various types of interleavers are well known in the art including, for example, diagonal interleavers, convolutional interleavers, block interleavers, inter-block interleavers, and pseudo-random interleavers. By using an interleaver with a pseudo-random structure, better turbo coding performance can be achieved.
turbo coding represents a significant improvement in the field of Forward Error Correction (FEC). There are many different turbo codes, but most types of turbo codes use multiple encoding steps separated by interleaving steps combined with the use of iterative decoding. This combination provides previously unavailable performance with respect to noise margin in the communication system. That is, turbo coding allows reliable communication at a lower power spectral density per noise per bit energy than was previously possible using existing forward error correction techniques.
For multi-component coding, such as turbo coding, optimal decoding is often a very complex task and may require a large number of time periods that are typically not available for instantaneous decoding. In practice, the task is almost impossible, theoretically requiring a completion time equivalent to the age of the universe. Iterative decoding techniques have been developed to overcome this problem. Rather than immediately determining whether a received bit is a 0 or a 1, the receiver assigns a value to each bit on a multi-level scale that represents the probability that the bit is a 1. A common scale, called log-likelihood ratio (LLR) probability, represents each bit as an integer in a certain range, e.g., { -32, 31 }. A value of 31 indicates that the transmitted bit has a high probability of being 0 and a value of-32 indicates that the transmitted bit has a high probability of being 1. A value of 0 indicates that the logical bit value is indeterminate.
Data represented on a multi-level scale is referred to as "soft data" and iterative decoding is typically soft-in/soft-out, i.e., the decoding process receives an input sequence of probabilities corresponding to bit values and provides corrected probabilities as output, taking into account the constraints of the encoding. In general, a decoder performing iterative decoding uses soft data from a previous iteration to decode soft data read by the receiver. During iterative decoding of multi-component codes, the decoder uses decoding from one code to improve decoding of the 2 nd code. When parallel encoders are used, as in turbo encoding, two corresponding decoders may be conveniently used for this purpose. Such iterative decoding is performed with multiple iterations until the soft data is believed to closely represent the transmitted data. Those bits that have a probability indicating that they are closer to 0 (e.g., a value between 0 and 31 on the scale described above) are assigned a binary 0, while the remaining bits are assigned a binary 1.
The LLR is thus a probability metric used by the turbo decoder to determine whether a given symbol was transmitted given a certain received symbol. To calculate the LLRs, accurate estimates of the SNR and channel coefficients (complex scale factors applied by the channel to the transmitted signal) are required. Accurate LLR values are particularly important in turbo decoding applications, where the LLR input is typically subjected to nonlinear operations, which can amplify inaccuracies in the LLR values and result in unacceptable decoder performance.
The LLR metrics for dense signal constellations are computationally intensive with the required high degree of accuracy. Either elaborate computational schemes have to be used or complex decoder structures have to be used. turbo encoders have long been significantly more complex to implement than convolutional encoders. Therefore, it has proven advantageous to maintain the turbo decoder structure at the cost of using computationally intensive LLR computation techniques. However, the use of computationally complex LLR calculation algorithms is undesirable due to the associated consumption in processor and memory resources. Moreover, computing the LLR metrics is particularly complex for certain modulation schemes such as 8-phase shift keying (8PSK) and 16-quadrature amplitude modulation (16 QAM). It would be desirable to provide a simplified method to derive the LLR metrics from an estimate rather than a direct calculation. Thus, there is a need for a simplified technique for computing soft decision input metrics for a turbo decoder without degrading the performance of the decoder.
Disclosure of Invention
The present invention is directed to simplified techniques for computing soft decision input metrics for a turbo decoder without degrading the performance of the decoder. Accordingly, in one aspect of the present invention, a method is provided for approximating log-likelihood ratio metrics for a plurality of turbo encoded symbols that have been modulated by a square quadrature amplitude modulation signal constellation having gray code labeling. The method advantageously comprises the steps of: extracting complex modulation symbol soft decisions on modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; scaling the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the most significant code symbols of the modulation symbols; and applying a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for remaining code symbols of the modulation symbols.
In another aspect of the invention, a receiver is provided that is configured to approximate log-likelihood ratio metrics for a plurality of turbo encoded symbols that have been modulated by square quadrature amplitude modulation signal constellations having gray code labeling. The receiver advantageously comprises: a demodulator configured to extract complex modulation symbol soft decisions on received modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; and a log-likelihood ratio calculation module coupled to the demodulator and configured to receive the complex modulation symbol soft decisions from the demodulator, scale the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for most significant code symbols of the modulation symbols, and apply a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for remaining code symbols of the modulation symbols.
In another aspect of the invention, a receiver is provided that is configured to approximate log-likelihood ratio metrics for a plurality of turbo encoded symbols that have been modulated by square quadrature amplitude modulation signal constellations having gray code labeling. The receiver advantageously comprises: means for extracting complex modulation symbol soft decisions on received modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; means for scaling the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the most significant code symbols of the modulation symbols; and means for applying a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the remaining code symbols of the modulation symbols.
In another aspect of the invention, a receiver is provided that is configured to approximate log-likelihood ratio metrics for a plurality of turbo encoded symbols that have been modulated by square quadrature amplitude modulation signal constellations having gray code labeling. The receiver advantageously comprises: a processor; and a processor-readable storage medium coupled to the processor and containing a set of instructions executable by the processor for extracting complex modulation symbol soft decisions on received modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; scaling the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for a most significant code symbol of the modulation symbols; and applying a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for remaining code symbols of the modulation symbols.
In another aspect of the invention, a method of approximating log-likelihood ratio metrics for a plurality of turbo encoded symbols that have been modulated by an M-phase shift keying signal constellation having gray code labeling is provided. The method advantageously comprises the steps of: extracting complex modulation symbol soft decisions on modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; scaling the orthogonal components to obtain log-likelihood ratio metrics for a most significant code symbol of the modulation symbols; scaling the in-phase component to obtain log-likelihood ratio metrics for a 2 nd most significant code symbol of the modulation symbol; and applying a product of a 1 st digit and a 2 nd digit to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for remaining code symbols of the modulation symbols, the 1 st digit depending on a size of the complex modulation symbol soft decisions and the 2 nd digit depending on a phase of the complex modulation symbol soft decisions.
In another aspect of the invention, a receiver is provided that is configured to approximate log-likelihood ratio metrics for a plurality of turbo encoded symbols that have been modulated by M-phase shift keying signal constellations having gray code labeling. The receiver advantageously comprises: a demodulator configured to extract complex modulation symbol soft decisions on received modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; and a log-likelihood ratio calculation module coupled to the demodulator and configured to receive the complex modulation symbol soft decisions from the demodulator, scale the quadrature component to obtain log-likelihood ratio metrics for a most significant code symbol of the modulation symbols, scale the in-phase component to obtain log-likelihood ratio metrics for a 2 nd most significant code symbol of the modulation symbols, and apply a product of a 1 st digit and a 2 nd digit to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for remaining code symbols of the modulation symbols, the 1 st digit depending on a size of the complex modulation symbol soft decisions and the 2 nd digit depending on a phase of the complex modulation symbol soft decisions.
In another aspect of the invention, a receiver is provided that is configured to approximate log-likelihood ratio metrics for a plurality of turbo encoded symbols that have been modulated by M-phase shift keying signal constellations having gray code labeling. The receiver advantageously comprises: means for extracting complex modulation symbol soft decisions on received modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; means for scaling the orthogonal components to obtain log-likelihood ratio metrics for a most significant code symbol of the modulation symbols; means for scaling the in-phase component to obtain a log-likelihood ratio metric for a 2 nd most significant code symbol of the modulation symbols; and means for applying a product of a 1 st digit and a 2 nd digit to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for remaining code symbols of the modulation symbols, the 1 st digit depending on a size of the complex modulation symbol soft decisions and the 2 nd digit depending on a phase of the complex modulation symbol soft decisions.
In another aspect of the invention, a receiver is provided that is configured to approximate log-likelihood ratio metrics for a plurality of turbo encoded symbols that have been modulated by M-phase shift keying signal constellations having gray code labeling. The receiver advantageously comprises: a processor; and a processor-readable storage medium coupled to the processor and containing a set of instructions executable by the processor for extracting complex modulation symbol soft decisions on received modulation symbols, wherein the modulation symbols are associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; scaling the orthogonal components to obtain log-likelihood ratio metrics for a most significant code symbol of the modulation symbols; scaling the in-phase component to obtain log-likelihood ratio metrics for a 2 nd most significant code symbol of the modulation symbol; and means for applying a product of a 1 st digit and a 2 nd digit to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the remaining code symbols of the modulation symbols, the 1 st digit being dependent on the magnitude of the complex modulation symbol soft decisions and the 2 nd digit being dependent on the phase of the complex modulation symbol soft decisions.
In another aspect of the invention, an apparatus is provided for estimating log-likelihood ratio (LLR) decoder metrics from soft decisions having an in-phase component and a quadrature component, the soft decisions having been constellation-modulated by an 8-phase shift keying (8PSK) signal having gray-coded symbols. The apparatus advantageously comprises: a 1 st multiplier configured to multiply the in-phase component by a 1 st constant value to produce a 1 st LLR metric; a 2 nd multiplier configured to multiply the quadrature component by a 1 st constant value to produce a 2 nd LLR metric; and a module configured to subtract the absolute value of the quadrature component from the absolute value of the in-phase component to (1) produce a difference, (2) add the absolute value of the in-phase component to the absolute value of the quadrature component to produce a sum, (3) divide the 2 nd constant value by the square of the sum of the squares of the in-phase component and the quadrature component to produce a quotient, and (4) multiply the difference, sum, and quotient to produce a 3 rd LLR metric.
In another aspect of the invention, a method is provided for estimating log-likelihood ratio (LLR) decoder metrics from soft decisions having an in-phase component and a quadrature component, the soft decisions having been constellation-modulated by an 8-phase shift keying (8PSK) signal having gray-coded symbols. The method advantageously comprises the steps of: multiplying the in-phase component by a 1 st constant value to produce a 1 st LLR metric; multiplying the quadrature component by a 1 st constant value to produce a 2 nd LLR metric; multiplying a subtraction value equal to a difference between the absolute value of the in-phase component and the absolute value of the quadrature component by an addition value equal to a sum of the absolute value of the in-phase component and the absolute value of the quadrature component to obtain an intermediate value; multiplying the intermediate value by a division value equal to a quotient of the 2 nd constant value and a square root of a sum of squares of the in-phase component and the quadrature component to obtain a 3 rd LLR metric.
In another aspect of the invention, an apparatus is provided for estimating log-likelihood ratio (LLR) decoder metrics from soft decisions having an in-phase component and a quadrature component, the soft decisions having been constellation-modulated by an 8-phase shift keying (8PSK) signal having gray-coded symbols. The apparatus advantageously comprises: means for multiplying the in-phase component by a 1 st constant value to produce a 1 st LLR metric; means for multiplying the orthogonal components by a 1 st constant value to produce a 2 nd LLR metric; means for multiplying an addition value equal to a sum of the absolute value of the in-phase component and the absolute value of the quadrature component by a subtraction value equal to a difference between the absolute value of the in-phase component and the absolute value of the quadrature component to obtain an intermediate value; means for multiplying the intermediate value by a division value equal to the 2 nd constant value and the quotient of the square root of the sum of the squares of the inphase component and the quadrature component to obtain a 3 rd LLR metric.
In another aspect of the invention, an apparatus is provided for estimating log-likelihood ratio (LLR) decoder metrics from soft decisions having an in-phase component and a quadrature component, the soft decisions having been constellation-modulated by an 8-phase shift keying (8PSK) signal having gray-coded symbols. The apparatus advantageously comprises a processor; and a storage element coupled to the processor and containing a set of instructions executable by the processor to multiply the in-phase component by a 1 st constant value to obtain a 1 st LLR metric, multiply the quadrature component by a 1 st constant value to obtain a 2 nd LLR metric, multiply a subtraction value equal to a difference between the absolute value of the in-phase component and the absolute value of the quadrature component by an addition value equal to a sum of the absolute value of the in-phase component and the absolute value of the quadrature component to obtain an intermediate value, and multiply the intermediate value by a division value equal to a quotient of the 2 nd constant value and a square root of a sum of squares of the in-phase component and the quadrature component to obtain a 3 rd LLR metric.
In another aspect of the invention, an apparatus is provided for estimating log-likelihood ratio (LLR) decoder metrics from soft decisions having an in-phase component and a quadrature component, the soft decisions having been modulated by a 16 quadrature amplitude modulation (16QAM) signal constellation having gray coded symbols. The apparatus advantageously comprises: a 1 st multiplier configured to multiply the in-phase component by a 1 st constant value to produce a 1 st LLR metric; a 2 nd multiplier configured to multiply the quadrature component by a 1 st constant value to produce a 2 nd LLR metric; a 1 st module configured to subtract a product of a carrier-to-signal-to-interference ratio (C/I) and a 2 nd constant value from an absolute value of the 2 nd LLR metric to generate a 3 rd LLR metric; and a 2 nd module configured to subtract a product of the C/I ratio and the 2 nd constant value from the absolute value of the 1 st LLR metric to generate a 4 th LLR metric.
In another aspect of the invention, a method is provided for estimating log-likelihood ratio (LLR) decoder metrics from soft decisions having an in-phase component and a quadrature component, the soft decisions having been modulated by a 16 quadrature amplitude modulation (16QAM) signal constellation having gray coded symbols. The method advantageously comprises the steps of: multiplying the in-phase component by a 1 st constant value to obtain a 1 st LLR metric; multiplying the orthogonal components by a 1 st constant value to obtain a 2 nd LLR metric; subtracting a product of a carrier-to-signal-to-interference ratio (C/I) and a 2 nd constant value from the absolute value of the 2 nd LLR metric to obtain a 3 rd LLR metric; and subtracting the product of the C/I ratio and the 2 nd constant value from the absolute value of the 1 st LLR metric to obtain a 4 th LLR metric.
In another aspect of the invention, an apparatus is provided for estimating log-likelihood ratio (LLR) decoder metrics from soft decisions having an in-phase component and a quadrature component, the soft decisions having been modulated by a 16 quadrature amplitude modulation (16QAM) signal constellation having gray coded symbols. The apparatus advantageously comprises: means for multiplying the in-phase component by a 1 st constant value to obtain a 1 st LLR metric; means for multiplying the orthogonal components by a 1 st constant value to obtain a 2 nd LLR metric; means for subtracting a product of a carrier-to-signal-to-interference ratio (C/I) and a 2 nd constant value from an absolute value of the 2 nd LLR metric to obtain a 3 rd LLR metric; and means for subtracting a product of the C/I ratio and the 2 nd constant value from the absolute value of the 1 st LLR metric to obtain a 4 th LLR metric.
In another aspect of the invention, an apparatus is provided for estimating log-likelihood ratio (LLR) decoder metrics from soft decisions having an in-phase component and a quadrature component, the soft decisions having been modulated by a 16 quadrature amplitude modulation (16QAM) signal constellation having gray coded symbols. The apparatus advantageously comprises a processor; and a memory element coupled to the processor and comprising a set of instructions executable by the processor to multiply the in-phase component by a 1 st constant value to obtain a 1 st LLR metric, multiply the quadrature component by a 1 st constant value to obtain a 2 nd LLR metric, subtract a product of a carrier-to-signal-to-interference ratio (C/I) and a 2 nd constant value from an absolute value of the 2 nd LLR metric to obtain a 3 rd LLR metric, and subtract a product of a C/I ratio and the 2 nd constant value from an absolute value of the 1 st LLR metric to obtain a 4 th LLR metric.
Drawings
Fig. 1 is a block diagram of a Code Division Multiplexing (CDM) transmitter.
Fig. 2 is a block diagram of a CDM receiver.
Fig. 3 is a block diagram of a Time Division Multiplexing (TDM) transmitter.
Fig. 4 is a block diagram of a TDM receiver.
Fig. 5 is a block diagram of circuitry for calculating a carrier-to-signal-to-interference ratio (C/I) for use with the forward link and the receiver of fig. 2 or 4.
Fig. 6 is a block diagram of an LLR estimation circuit usable in the circuit of fig. 5 with respect to an 8PSK modulation scheme.
Fig. 7 is a block diagram of an LLR estimation circuit usable in the circuit of fig. 5 with respect to a 16QAM modulation scheme.
Fig. 8 is a block diagram of a communication system, model.
Fig. 9 is a diagram of a 4qam (qpsk) signal constellation mapped with gray coded symbols.
Figure 10 is a diagram of a 16QAM signal constellation mapped with gray coded symbols.
Fig. 11 is a diagram of a 64QAM signal constellation mapped with gray coded symbols.
Fig. 12 is a diagram of 8PSK signal constellation mapped with gray coded symbols.
Fig. 13 is a diagram of a 16PSK signal constellation mapped with gray coded symbols.
Fig. 14 is a graph of 16QAM coded symbol LLR versus modulation symbol metrics.
Fig. 15 is a graph of 64QAM coded symbol LLRs versus modulation symbol metrics.
Fig. 16 is a graph of 256QAM coded symbol LLRs versus modulation symbol metrics.
Fig. 17 is a graph of 8PSK encoded symbol LLR versus modulation symbol metrics.
Fig. 18 is a graph of 16PSK encoded symbol LLR versus modulation symbol metrics.
Detailed Description
A Code Division Multiplexed (CDM) transmitter 10 constructed in accordance with one embodiment is shown in fig. 1. Various details of the transmitter, such as timing circuits, filters, and amplifiers, are omitted from the figure for clarity. The omitted circuit can be easily constructed and implemented by those of ordinary skill in the art.
The transmitter 10 comprises a computer 12, the computer 12 comprising transmitter software executed via a baseband processor (not shown) in the computer 12. The computer 12 is coupled to a turbo encoder 14 and a time division combiner 16. The turbo encoder is coupled to a channel interleaver 18 which is coupled to the 1 st input of a 1 st multiplier 20. A 1 st Walsh function generator 22 is coupled to the 2 nd input of the 1 st multiplier 20. The output of the 1 st multiplier 20 is coupled to the 1 st input of a combiner 24.
The output of the time division combiner 16 is coupled to the 1 st input of a 2 nd multiplier 26, the 2 nd input of which is coupled to a 2 nd Walsh function generator 28. The output of the 2 nd multiplier 26 is coupled to the 2 nd input of the combiner 24. The output of the combiner 24 is coupled to a quadrature pseudo-noise (PN) sequence spreader 30. The output of the PN spreader is input to a modulator 32 coupled to an antenna 34.
In operation, a signal containing data, such as voice data or other file data, is passed from the computer 12 to the turbo encoder 14. The turbo encoder 14 encodes the data signal. The turbo encoder 14 is a standard turbo encoder and operates according to turbo encoding principles well known in the art. In one embodiment, the turbo encoders 14 are cascaded with an interleaver (not shown) disposed between inner and outer constituent convolutional encoders (not shown). In another embodiment, the turbo encoder 14 is a parallel turbo encoder designed according to principles well known in the relevant art.
The encoded data signal output from the turbo encoder 14 is then interleaved by a channel interleaver 18 in preparation for Walsh encoding, Pseudo Noise (PN) spreading, and modulation. The channel interleaver 18 may be implemented by a conventional interleaver such as a block interleaver.
The computer 12 also provides a predetermined pilot signal, which in this detailed embodiment is a constant equal to 1, and a control signal to the time division combiner. The control signals contain rate control or power control information for communication to a corresponding receiver (as described in detail below) to facilitate power and/or code rate control to maximize efficiency and throughput of the communication system.
The time division combiner 16 mixes the control signal with the pilot signal according to a conventional time division combining method. The combined signal is input to the 2 nd multiplier 26 where it is multiplied by a predetermined Walsh function provided via the 2 nd Walsh function generator 28. Similarly, the interleaved data signal output from the channel interleaver 18 is provided to the 1 st multiplier 20 where it is multiplied by a predetermined Walsh function provided by the 1 st Walsh function generator 22.
The resulting Walsh coded outputs from the 1 st multiplier 20 and the 2 nd multiplier 26 are combined by the combiner 24, spread by the PN spreader 30, and then modulated and converted to radio frequency by the modulator 32 in preparation for transmission over a channel via the antenna 34.
The resulting signal transmitted via the antenna 34 is a composite signal having a data signal, a pilot signal, and a control signal. Once broadcast over the channel, the composite signal will experience multipath fading and channel interference, which must be effectively detected and compensated for by the receiver system receiving the transmitted signal.
Those of ordinary skill in the art will appreciate that the Walsh functions provided by the 1 st Walsh function generator 22 and the 2 nd Walsh function generator 28 may be replaced with a PN function generator or a combination of a Walsh function generator and a PN function generator. Furthermore, the transmitter 10 may be implemented in a base station and/or a mobile station in a cellular or PCS communication system.
In this detailed description, the terms signal-to-interference ratio and signal-to-noise ratio are equivalent terms.
A CDM receiver 40 constructed in accordance with one embodiment for use with the CDM transmitter of fig. 1 is shown in fig. 2. The receiver 40 includes a receiver antenna 42 coupled to demodulation circuitry 44. The demodulation circuit 44 is coupled to an Automatic Gain Control (AGC) circuit 46 that is coupled to an analog-to-digital converter (ADC) 48. The output of the ADC 48 is coupled to the input of the 1 st receiver multiplier 50. The output of the ADC 48, representing digital samples, is also provided as an input to carrier-to-signal-to-interference ratio (C/I) estimation and log-likelihood ratio (LLR) calculation circuitry, as described in detail below.
The other input of the 1 st receiver multiplier 50 is coupled to the output of a PN sequence generator 52. The output of the 1 st receiver multiplier 50 is coupled in parallel to the inputs of a 2 nd receiver multiplier 54 and a 3 rd receiver multiplier 56. The 1 st receiver Walsh generator circuit 58 and the 2 nd receiver Walsh generator circuit 60 also provide inputs to the 2 nd receiver multiplier 54 and the 3 rd receiver multiplier 56, respectively. The outputs of the 2 nd receiver multiplier 54 and the 3 rd receiver multiplier 56 are coupled to the inputs of a 1 st accumulator 62 and a 2 nd accumulator 64, respectively. The output of the 1 st accumulator 62 is coupled to a sample splitter and despreader that provide outputs to a C/I estimation circuit and an LLR calculation circuit, as described in detail below.
In operation, signals transmitted over a channel, such as RF signals transmitted by the transmitter 10 of FIG. 1, are received by the antenna 42 of the receiver 40. The received RF signal is converted into an intermediate frequency signal by the demodulator 44 and then converted into a baseband signal. The gain of the baseband signal is adjusted by the AGC circuit 46 and then converted to a digital signal by the ADC 48. The baseband signal is then multiplied by a PN sequence related to the PN sequence used in PN spreader 30 of fig. 1, via the PN sequence generator 52 and the 1 st receiver multiplier 50. In this detailed embodiment, the PN sequence and its inverse are the same because with binary operation (in GF 2), the inverse of 1 is 1 and the inverse of 0 is 0.
The 1 st receiver multiplier 50 then outputs a partially despread signal, splitting the signal into two separate paths. On one path, the 2 nd receiver multiplier 54 multiplies the partially spread sequence with the Walsh function provided by the 1 st receiver Walsh function generator 58. The Walsh functions provided are related to the Walsh functions provided by the 1 st Walsh function generator 22 of fig. 1. The resulting despread signal samples are input to the 1 st accumulator 62 where the signal samples are accumulated over a predetermined number of samples. The accumulated despread data samples are provided to the sample separator 66. The sample separator 66 outputs the pilot signal and the control signal extracted from the despread signal to a C/I estimation circuit and an LLR calculation circuit, as described in detail below.
Similarly, the despread signal sample output from the 3 rd receiver multiplier 56 is accumulated by the 2 nd accumulator, which outputs a data signal comprising data signal samples to the C/I estimation circuit and LLR circuit, as described in detail below.
A Time Division Multiplexed (TDM) transmitter 70 constructed according to one embodiment is shown in fig. 3. The TDM transmitter 70 is advantageously similar to the CDM transmitter 10 of fig. 1, except that the time division combiner 16, the multipliers 20, 26, the Walsh function generators 22, 28, and the summer 24 of fig. 1 are replaced by a time division combiner 72.
A TDM receiver 80 constructed in accordance with one embodiment is shown in fig. 4. The TDM receiver 80 is advantageously similar to the CDM receiver 40 of fig. 1, except that the multipliers 54, 56, the Walsh function generators 58, 60, the accumulators 62, 64, and the sample separator 66 of fig. 2 are replaced by an accumulator 82 and a TDM sample separator 84. The accumulator 82 receives the digital spread samples from the multiplier 50, accumulates the samples, and then provides the accumulated samples to the TDM sample separator 84. The TDM sample splitter 84 extracts data samples, pilot samples, and control samples from the accumulated and despread digital signal. The data samples, pilot samples, control samples, and the digital sample output from ADC 48 are provided to a C/I estimation and LLR circuit, as described in detail below.
A C/I estimation circuit 120 constructed in accordance with one embodiment is shown in fig. 5. The C/I estimation circuit 120 is advantageously adapted for use with the forward link and the receiver of fig. 2 or 4. The C/I estimation circuit 120 includes a PN despreader 122 that can replace the multiplier 50, the PN sequence generator 52, and the accumulator 82 of the receiver 80 of fig. 4. The M Walsh decover circuit 124 may replace the TDM sample splitter 84 of fig. 4.
The C/I estimation circuit 120 includes the PN despreader 122, an M Walsh decover circuit 124, a received signal total energy (Io) calculation circuit 126, a 1 st constant circuit 136, a pilot filter 128, a subtractor 132, a 1 st multiplier 134, a pilot energy calculation circuit 138. A look-up table (LUT)140, a 2 nd multiplier 142, and a C/I accumulation circuit 144. In the C/I estimation circuit 120, the PN despreader 122 receives the digital in-phase (I) and quadrature (Q) signal outputs from the ADC 48 of fig. 4. The PN despreader 122 provides the M Walsh decover circuit 124 and the I in paralleloThe calculation circuit 126 provides an input. The M Walsh decover circuit 124 provides an input to the pilot filter 128 and to the constant divider circuit 130 in the path weight and combining circuit 158.
The output of the energy calculation circuit 126 is coupled to the positive terminal of the subtractor circuit 132. The negative terminal of the subtractor circuit 132 is coupled to the output terminal of the 1 st multiplier 134. The 1 st input of the 1 st multiplier 134 is coupled to the output of the 1 st constant circuit 136. The 2 nd input of the 1 st multiplier 134 is coupled to the output of the pilot energy calculation circuit 138. The pilot filter 128 provides an input to the pilot energy calculation circuit 138.
The output of the subtractor 132 is coupled to the LUT 140. The output of the LUT 140 is coupled in parallel to the 1 st input of the 2 nd multiplier 142 and the 1 st input of the 3 rd multiplier 146 in the path weighting and combining circuit 158. The 2 nd input of the 2 nd multiplier 142 is coupled to the output of the 1 st multiplier 134. The output of the 2 nd multiplier 142 is coupled to the C/I summation circuit 144, the output of which provides an input to the LLR circuit 96.
The path weighting and combining circuit 158 includes a 2 nd constant generating circuit 150, a 4 th multiplier 148, the 3 rd multiplier 146, the constant divider circuit 130, a complex conjugate circuit 152, a 5 th multiplier 154, and a path accumulator circuit 156. In the path weighting and combining circuit 158, the 1 st terminal of the 4 th multiplier 148 is coupled to the output of the pilot filter 128, which is also coupled to the input of the pilot energy calculation circuit 138 in the C/I estimation circuit 120. The 2 nd terminal of the 4 th multiplier 148 is coupled to the 2 nd constant generating circuit 150. The output of the 4 th multiplier 148 is coupled to the 2 nd input of the 3 rd multiplier 146. The output of the 3 rd multiplier 146 provides an input to the complex conjugate circuit 152. The output of the complex conjugate circuit 152 is coupled to the 1 st input of the 5 th multiplier 154. The output of the constant divider 130 is coupled to the 2 nd input of the 5 th multiplier 154. The output of the 5 th multiplier 154 is coupled to the input of the path accumulator 156. The output of the path accumulator 156 is coupled to the 2 nd input of the LLR circuit 96. The output of the second LLR circuit is coupled to an input of a decoder (not shown).
In operation, the PN despreader 122 receives the I and Q signals and despreads the L fingers, the Lth individual lane (1). The PN despreader 122 despreads the I and Q signals using an inverse of the PN sequence used to spread the I and Q signals prior to transmission over a channel. The construction and operation of the PN despreader 122 is well known in the art.
The despread signal is output from the PN despreader 122 and input to the M Walsh decover circuit 124 and the IoA calculation circuit 126. Said IoThe calculation circuit 126 calculates the total energy (I) received for each symbolo) Which includes a desired signal component and interference and noise components. Said IoThe calculation circuit provides I according to the following equationoIs estimated by
img id="idf0002" file="A20071013644700182.GIF" wi="343" he="42" img-content="drawing" img-format="GIF"/
Where N is the number of symbols per pilot burst, and in this detailed embodiment is 64, and · represents the received despread signal output from the PN despreader 122.
Those of ordinary skill in the art will appreciate that I may be calculated prior to despreading by the PN despreading 122 in alternative embodimentso. For example, the said IoThe calculation circuit 126 may receive inputs directly from the I and Q signals received from the ADC 48 of fig. 2 and 4, rather than receiving an input provided by the PN despreader 122, in this case, at the IoI to be provided at the output of the calculation circuit 126oAnd (4) performing equivalent estimation.
The M Walsh decover circuit 124 decovers orthogonal data signals, referred to as data channels, and pilot signals, referred to as pilot channels, according to methods known in the art. In the present detailed embodiment, the orthogonal data signal corresponding to one data channel s is represented by the following equation:
img id="idf0003" file="A20071013644700183.GIF" wi="374" he="27" img-content="drawing" img-format="GIF"/
where M is the number of symbols per Walsh symbol,is the modulation symbol energy of the ith multipath component,is the phase of the data channel s, and XtIs the information-bearing component of the data channel s. The decovered data is provided to a decoder (as described in detail below) and the constant divider circuit 130 of the path weighting and combining circuit 158.
It should be understood that although the embodiments described herein are applicable to using signals that include various Walsh codes, those of ordinary skill in the art may readily adapt the embodiments for use with other types of codes.
The pilot channel is provided to a pilot filter 128. The pilot filter 128 is an averaging filter that acts as a low pass filter that removes higher frequency noise and interference components from the pilot channel. The output p of the pilot filter 128 is represented by the following equation:
img id="idf0006" file="A20071013644700186.GIF" wi="374" he="29" img-content="drawing" img-format="GIF"/
where M is the number of symbols per Walsh symbol,is the pilot symbol energy of the ith multipath component, and θlIs the phase of the filtered pilot channel p.
An energy estimate of the filtered pilot channel p, which is the square of the complex amplitude of the filtered pilot channel p represented by equation (3), is calculated by the pilot energy calculation circuit 138. Multiplying the square of the complex amplitude of the filtered pilot channel p by a predetermined scaling factor c described by the following equation:
img id="idf0008" file="A20071013644700191.GIF" wi="374" he="44" img-content="drawing" img-format="GIF"/
wherein IorIs the received energy of the desired signal, i.e. is equivalent to I minus the noise and interference componentso。EpIs the pilot symbol energy. In many wireless communication systems, the scaling factor c is a known forward link constant.
The energy of the filtered pilot signal p is multiplied by the scaling factor c by the 1 st multiplier 134 to produce the energy of the received desired signal associated with the l multipath component of the received I and Q signals (I minus the noise and interference components)o) Is precisely estimated
From the I by the subtractor 132oSubtracting the accurate estimate from the estimate ofTo generate an interference energy (N) associated with the l-th multipath componentt，l) Accurate measurement of. Then adding Nt，lIs provided to the LUT 140 which weights the 3 rd product in the path weighting and combining circuit 1581 st input/output N of the multiplier 146 and the 2 nd multiplier 142t，lThe reciprocal of (c). Coupling the 2 nd input of the 2 nd multiplier 142 to the output of the 1 st multiplier 134, which is provided at the 2 nd input terminal of the 2 nd multiplier 142The 2 nd multiplier 142 outputs C/I according to the following equationl(C/I associated with the l-th multipath component):
img id="idf0012" file="A20071013644700195.GIF" wi="374" he="49" img-content="drawing" img-format="GIF"/
the precise C/I values are then accumulated over L passes in the received signal by the C/I accumulator circuit 144. The accumulated C/I values are then provided to the LLR circuit 96 and a rate/power request generation circuit (not shown), the construction of which is known in the art.
In the path weighting and combining circuit 158, the 4 th multiplier 148 multiplies the filtered pilot signal p by the constant k provided by the 2 nd constant generating circuit 150. The constant k is calculated according to the following equation:
img id="idf0013" file="A20071013644700196.GIF" wi="374" he="51" img-content="drawing" img-format="GIF"/
wherein EsIs the modulation symbol energy, EpIs the pilot symbol energy, and M is the number of Walsh symbols per symbol as described above. For reverse link and forward link transmissions, the EsAnd EpThe ratio of (c) is often known or can be determined.
The output of the 4 th multiplier 148 provides the channel coefficient described by the equation belowEstimation of (2):
img id="idf0015" file="A20071013644700201.GIF" wi="374" he="28" img-content="drawing" img-format="GIF"/
whereinIs an estimate of the modulation symbol energy of the l-th multipath component, anIs an estimate of the phase of the pilot signal.
The interference energy N associated with the l-th multipath component is then used by the 3 rd multiplier 146t，lIs multiplied by the reciprocal ofAnd estimating the channel. The interference energy Nt，lIncluding interference and noise components. The complex conjugate circuit 152 then calculates the conjugate of the output of the 3 rd multiplier 146, which represents the path combining weight of the largest ratio. The corresponding data symbol output from the divider circuit 130 is then multiplied by the maximum ratio path combining weight by the 5 th multiplier 154. The data symbol d is represented by the following equation:
img id="idf0018" file="A20071013644700204.GIF" wi="406" he="29" img-content="drawing" img-format="GIF"/
the variables are the same as given in equation (2).
The output of the 5 th divider 154 represents the optimally weighted data signal, which is then accumulated by the lane combine circuit 156 over the L lanes containing the signal. The resulting optimally combined data signal is provided to the LLR circuit 96, which facilitates the computation of the optimal soft decoder input for the decoder (as described in detail below).
Can be used (Y)I，YQ) To indicate soft decisions on modulation symbols produced by path combining and repetition combining, where YIRepresenting soft decisions in phase, and YQRepresenting orthogonal soft decisions. The soft decision (Y) is passed by the LLR circuit 96I，YQ) Into LLRs on the code symbols. The LLRs constitute the soft-decision input to a turbo decoder (not shown). Such as those of ordinary skill in the artIt will be appreciated that the soft decision (Y) is indicated by the modulation scheme usedI，YQ) The manner of transformation into LLR metrics on the underlying code symbols and the construction of LLR circuit 96.
Those of ordinary skill in the art will appreciate that the computation of LLR metrics in Quadrature Phase Shift Keying (QPSK) modulation schemes is relatively straightforward compared to the computation of modulation schemes such as 8 phase shift keying (8PSK) and 16 quadrature amplitude modulation (16 QAM). In QPSK modulation scheme, each complex modulation symbol is soft-decided (Y)I，YQ) Conveying information about two coded symbols c1And c0The information of (1). In fact, the in-phase component YICarrying information about coded symbols c0All information of (2), and the orthogonal component YQCarrying the code word c in respect of the remainder1All of the information of (a). If in this way (Y) is to beI，YQ) Normalized, i.e. its average amplitude is equal to said C/I ratio, the modulation soft decision Y can be made according to the following equationIAnd YQInto said coded symbols c0And c1LLR metric above:
LLR(c0)＝2*YI [9]
and
LLR(c1)＝2*YQ [10]
exemplary circuits FOR calculating LLR metrics FOR QPSK modulation schemes are described IN U.S. patent application Ser. No. 09/311,793, filed on 25.5.13.1999, entitled "SYSTEM AND METHOD FOR PERFORMING APPARATUS DEMODULATION OF THE INVENTION-ENCODED SIGNALS VIA P-ASSISTED COHERENT DEMODULATION" (U.S. patent number 6377607, published on 23.4.2002), and U.S. patent application Ser. No. 09/310,053, 11.1999, filed on 11.5.11. SYSTEM AND METHOD FOR PROVIDING AN ACCURATEMATION OF RECEIVED SIGNAL INTERFERENCE FOR USE IN WIRELESS COMMUNICATION SYSTEMS ", both OF which are assigned to the assignee OF the present invention and incorporated herein by reference.
Computing the LLR metrics becomes much more difficult if an 8PSK modulation scheme is used. According to one embodiment, LLR metrics used with respect to an 8PSK modulation scheme with gray coded symbols are advantageously estimated. As described above, (Y) is substituted in this mannerI，YQ) Normalized, i.e. its average amplitude is equal to the C/I ratio. Since the modulation scheme is 8PSK, each complex modulation symbol is soft-decided (Y)I，YQ) Conveying information about 3 code symbols c2、c1And c0The information of (1). 1 st coded symbol c2Is the most significant symbol of the received 8PSK modulated codeword conveying information about the transmitted bits. 3 rd code symbol c0Is the least significant symbol of the received 8PSK modulated codeword. Soft decision (Y) from modulation symbols according to the following equationI，YQ) Advantageously obtaining said coded symbol c2、c1And c0Simplified estimation of the LLR metric on:
img id="idf0019" file="A20071013644700211.GIF" wi="479" he="51" img-content="drawing" img-format="GIF"/
LLR(c1)＝2.6131YI [12]
and
LLR(c2)＝2.6131YQ [13]
according to one embodiment, an LLR estimation circuit 200 may be used in place of the LLR circuit 96 of fig. 5 to provide a simplified estimate of LLR metrics for an 8PSK modulation scheme with gray code flags, as shown in fig. 6. The LLR estimation circuit 200 includes 1 st, 2 nd, 3 rd and 4 th multipliers 202, 204, 206, 208, an absolute value circuit 210, a squaring circuit 212, 1 st and 2 nd adders 214, 216, a subtractor 218 and a LUT 220. The absolute value circuit 210 is coupled to the squaring circuit 212, the 2 nd adder 216, and the subtractor 218. The squaring circuit 212 is coupled to the 1 st adder 214. The 1 st adder 214 is coupled to the LUT220, which is coupled to the 3 rd multiplier 206. The 3 rd multiplier 206 is coupled to the 4 th multiplier 208. The 2 nd adder 216 is coupled to the 3 rd multiplier 206 and the subtractor 218 is coupled to the 4 th multiplier 208.
In operation, the in-phase component Y of the demodulated soft decision is processedIIs provided to the 1 st multiplier 202. Constant digital value 2.6131 is also provided to 1 st multiplier 202. The 1 st multiplier 202 multiplies the in-phase component YIMultiplied by the digital value 2.6131 to produce a coded symbol c1LLR metric LLR (c) on1). Those of ordinary skill in the art will appreciate that the numerical value need not be limited exactly to 2.6131. Other values may be used to produce a slightly inaccurate LLR metric LLR (c)1) Is estimated.
Orthogonal component Y of demodulated soft decisionQTo the 2 nd multiplier 204. Constant digital value 2.6131 is also provided to multiplier 2 204. The 2 nd multiplier 204 multiplies the quadrature component YQMultiplied by the digital value 2.6131 to produce a coded symbol c2LLR metric LLR (c) on2). In the artThose of ordinary skill will appreciate that the numerical values need not be limited to 2.6131 precisely. Other values may be used to produce a slightly inaccurate LLR metric LLR (c)2) Is estimated.
Also the in-phase and quadrature components YIAnd YQAbsolute value Y provided for generating said componentII and YQAbsolute value circuit 210 of | is provided. The absolute value | YII and YQThe | is provided to a squaring circuit 212, which squares the provided value to produce a squared value YI 2And YQ 2. The absolute value circuit 210 and the squaring circuit 212 may be implemented with hardware circuits structured as understood by those of ordinary skill in the art. The absolute value circuit 210 and the squaring circuit 212 may also be implemented with a conventional processor or DSP executing a stored set of software or firmware instructions. Or a combination of the two implementations may be used. Squaring the value YI 2To the 1 st adder 214. The squared value Y is alsoQ 2To the 1 st adder 214. The 1 st adder 214 adds the two squared values and provides the sum (Y) to the LUT220I 2+YQ 2)。
The LUT220 is advantageously a ROM memory configured to store the quotient of 1.0824 and the square root of the values of the predetermined range of values. In alternative embodiments, the LUT220 may be implemented as any conventional form of non-volatile storage medium. A conventional processor or DSP (not shown) may be used to access the LUT220 and/or perform operations performed by other circuitry of the LLR estimation circuit 200. Those of ordinary skill will appreciate that numbers other than 1.0824 may be used, with the resulting result being a slightly inaccurate estimate of the LLR metric. For example, in one embodiment, the LUT220 stores the quotient of 1 and the square root of the values of a predetermined range of values.
LUT220 generates 1.0824 and the squared component Y for the 3 rd multiplier 206I 2And YQ 2The quotient of the square root of the sum of (c), the absolute value circuit 210 also compares the absolute value of the component | YII and YQThe 2 nd adder 216 is provided with. The 2 nd adder 216 adds the absolute value | YII and YQAdd | and sum | YI|+|YQIs provided to the 3 rd multiplier 206. The 3 rd multiplier 206 multiplies the sum YI|+|YQI and the quotientMultiplies and provides the resulting product to the 4 th multiplier 208.
The absolute value circuit 210 also compares the absolute value of the component | YII and YQAnd | is provided to subtractor 218. The subtractor 218 derives the absolute value | Y of the in-phase componentISubtracting absolute value Y of the orthogonal component from |QAnd will be different by YI|-|YQThe 4 th multiplier 208 is provided with. The 4 th multiplier 208 multiplies the difference YI|-|YQL is multiplied by the product provided by multiplier 3 206 to produce the code symbol c0LLR metric LLR (c) on0)。
Computing LLR metrics is also much more difficult than for QPSK modulation scheme if 16QAM modulation scheme is used. According to one embodiment, LLR metrics used with respect to a 16QAM modulation scheme with gray coded markers are advantageously estimated. As described above, (Y) is substituted in this mannerI，YQ) Normalized, i.e. its average amplitude is equal to the C/I ratio. Since the modulation scheme is 16QAM, each complex modulation symbol is soft-decided (Y)I，YQ) Conveying information about 4 code symbols c3、c2、c1And c0The information of (1). 1 st coded symbol c3Is the most significant symbol of a received 16QAM modulated codeword conveying information about the transmitted bits. 4 th coded symbol c0Is the least significant symbol of the received 16QAM modulated codeword. The in-phase component YICarrying information about coded symbol pairsc1And c0All information of (2), and the orthogonal component YQCarrying coded symbols c on the remainder3And c2All of the information of (a). Soft decision (Y) from modulation symbols according to the following equationI，YQ) Advantageously obtaining said coded symbol c3、c2、c1And c0Simplified estimation of the LLR metric on:
LLR(c0)＝1.2649|YI|-0.8(C/I) [14]
img id="idf0022" file="A20071013644700232.GIF" wi="447" he="62" img-content="drawing" img-format="GIF"/
LLR(c0)＝1.2649|YQ|-0.8(C/I) [16]
and
img id="idf0023" file="A20071013644700233.GIF" wi="448" he="63" img-content="drawing" img-format="GIF"/
equations (15) and (17) may be advantageously simplified by replacing the bracketed term in the equation with a value of 1, resulting in the following LLR (c)1) And LLR (c)3) The equation of (c):
LLR(c1)≈1.2649YI [18]
LLR(c3)≈1.2649YQ [19]
according to one embodiment, an LLR estimation circuit 300 may be used in place of LLR circuit 96 of fig. 5 to provide a simplified estimate of LLR metrics for a 16QAM modulation scheme with gray code flags, as shown in fig. 7. The LLR estimation circuit 300 includes 1 st, 2 nd and 3 rd multipliers 302, 304, 306, 1 st and 2 nd absolute value circuits 308, 310, and 1 st and 2 nd subtractors 312, 314. The 1 st multiplier 302 is coupled to the 2 nd absolute value circuit 310. The 2 nd multiplier 304 is coupled to the 1 st absolute value circuit 308. The 1 st absolute value circuit 308 is coupled to the 1 st subtractor 312. The 1 st subtractor 312 is also coupled to the 3 rd multiplier 306. The 2 nd absolute value circuit is coupled to the 2 nd subtractor 314. The 2 nd subtractor 314 is also coupled to the 3 rd multiplier 306.
In operation, the in-phase component Y of the demodulated soft decision is processedIIs provided to the 1 st multiplier 302. Constant digital value 1.2649 is also provided to 1 st multiplier 302. The 1 st multiplier 302 multiplies the in-phase component YIMultiplied by the digital value 1.2649 to produce a coded symbol c1LLR metric LLR (c) on1). Those of ordinary skill in the artThe skilled person will appreciate that the numerical values need not be limited to 1.2649 precisely. Other values may be used to generate the LLR metric LLR (c)1) Other estimates of (a).
Orthogonal component Y of demodulated soft decisionQIs provided to the 2 nd multiplier 304. Constant digital value 1.2649 is also provided to multiplier 2 304. The 2 nd multiplier 304 multiplies the quadrature component YQMultiplied by the digital value 1.2649 to produce a coded symbol c3LLR metric LLR (c) on3). Those of ordinary skill in the art will appreciate that the numerical value need not be limited exactly to 1.2649. Other values may be used to generate the LLR metric LLR (c)3) Other estimates of (a).
The product (i.e., LLR metric LLR (c)) output by the 1 st multiplier 3021) Is provided to the 2 nd absolute value circuit 310. The product (i.e., LLR metric LLR (c)) output by the 2 nd multiplier 3043) Is supplied to the 1 st absolute value circuit 308. The 1 st absolute value circuit 308 generates an absolute value of the product output by the 2 nd multiplier 304 and supplies the absolute value to the 1 st subtractor 312. The 2 nd absolute value circuit 310 generates an absolute value of the product output by the 1 st multiplier 302 and supplies the absolute value to the 2 nd subtractor 314. The 1 st and 2 nd absolute value circuits 308, 310 may be implemented with hardware circuits constructed as understood by those of ordinary skill in the art. The absolute value circuits 308, 310 may also be implemented with a conventional processor or DSP executing a stored set of software or firmware instructions. Or a combination of the two implementations may be used.
The estimate of the C/I ratio is provided to a 3 rd multiplier 306. A constant digital value of 0.8 is also provided to the 3 rd multiplier 306. The 3 rd multiplier 306 multiplies the C/I value by the digital value of 0.8. And provides the resulting product result to the 1 st and 2 nd subtractor circuits 312, 314. Those of ordinary skill in the art will appreciate that the numerical value need not be limited to exactly 0.8. Other values may be used to produce a slightly inaccurate estimate of the LLR metric.
The 1 st subtractor 312 subtracts the product output by the 3 rd multiplier 306 from the absolute value provided by the 1 st absolute value circuit 308 to produce a symbol c2LLR metric LLR (c) on2). The 2 nd subtractor 314 subtracts the product output by the 3 rd multiplier 306 from the absolute value provided by the 2 nd absolute value circuit 310 to produce a symbol c0LLR metric LLR (c) on0)。
By using a multi-level modulation scheme (such as QAM or MPSK) in conjunction with powerful coding techniques (such as turbo coding), reliable communication with high spectral efficiency is achieved. the turbo decoding algorithm uses soft-decision estimates output by the encoder to recover the encoded data. Most implementations of turbo decoders use LLRs on binary symbols at the encoder output as their soft-decision input. Extracting LLRs from the demodulator soft decisions is a computationally complex task, except for the simplest modulation schemes such as BPSK or QPSK. Thus, according to one embodiment, a simplified process is provided for approximate computation of LLRs from modulation symbol soft decisions for square QAM constellations (such as, for example, 16QAM, 64QAM, and 256QAM) and MPSK constellations (such as, for example, 8PSK and 16 PSK).
In a typical communication system 400 as illustrated in fig. 8, binary data d to be transmitted is turbo encoded by a turbo encoder 402 with turbo encodingnIs encoded, the turbo encoder generates a sequence of binary symbols b often referred to as code symbolsn. A number of code symbols are partitioned together and mapped to points on a signal constellation by a signal mapping module 404, thereby generating a complex sequence of modulation symbols xn. This sequence is applied to a modulator 406 that produces a continuous-time waveform that is transmitted over a channel 408.
At the receiver (not shown), the demodulator 410 uses the output to generate a complex soft decision sequence yn. Each soft decision represents an estimate of a modulation symbol transmitted on the channel 408. The estimate is used by LLR calculation module 412 to extract LLRs for the code symbols associated with the given modulation symbol. turbo decoder 414 uses codingA sequence of code symbol LLRs to decode the originally transmitted binary data.
In one embodiment, a square QAM constellation and rules for assigning a block of binary (coded) symbols to each point on the signal constellation are defined. Defining a square QAM constellation with an index m as having a value of 4mThe signal constellation of the dots, which can be regarded as having 2mCartesian products of the two Pulse Amplitude Modulated (PAM) constellations of a point. Each signal point is represented by its index (i, j), where 0 ≦ i, j < 2m. The position of the (i, j) th point on the signal constellation is given by:
ci，j＝{(2m-1-2i)Δ，(2m-1-2j) Δ } whereinimg id="idf0024" file="A20071013644700261.GIF" wi="135" he="49" img-content="drawing" img-format="GIF"/
The above definition ensures that the mean energy (i.e. the squared euclidean norm) of the signal constellation is normalized to 1. Table 1 below shows the values of the standard parameter Δ for various square QAM constellations.
TABLE 1 energy normalization for various square QAM constellations
Each signal point is marked with a binary string representing a block of coded symbol values associated with the modulation symbol. In a particular embodiment, gray code mapping is used to associate modulation symbols with blocks of code symbols. Gray code mapping is well known to those of ordinary skill in the art. It should be understood that other forms of mapping may be used. According to the gray-coded mapping, the label of the (i, j) th point on the constellation is given by:
li，j＝b2m-1b2m-2b2m-3...bm+2bm+1bm b2m-1b2m-2b2m-3...bm+2bm+1bmwherein b isk＝0，1； img id="idf0026" file="A20071013644700263.GIF" wi="142" he="43" img-content="drawing" img-format="GIF"/img id="idf0027" file="A20071013644700264.GIF" wi="163" he="43" img-content="drawing" img-format="GIF"/
The function gray (i) represents the well-known gray coding mapping, where
gray(0)＝0，gray(1)＝1，gray(2)＝(11)2＝3，gray(3)＝(10)2＝2，gray(4)＝(110)26, and so on. The gray code mapping may be formally defined as follows:
gray(0)＝0，gray(1)＝1，gray(k)＝2n-1+gray(2n-1-k), wherein 2n-1≤k＜2n。
Using the value bk(i, j) ═ 0, 1 to denote the label li，jThe kth component of (1).
The above signal constellation for m-1 is depicted in fig. 9. The above signal constellation for m-2 is depicted in fig. 10.
The above signal constellation for m-3 is depicted in fig. 11.
In one embodiment, a Multiple Phase Shift Keying (MPSK) signal constellation and a rule for assigning a block of binary (coded) symbols to each point on the signal constellation are defined. The MPSK constellation (also commonly referred to as 2) with index mm-PSK constellation) is defined as having a position 2 on the unit circlemSignal conformation of the dots. Each signal point is represented by its index i, where 0 ≦ i < 2m. The position of the ith point on the signal conformation is given by:
img id="idf0028" file="A20071013644700271.GIF" wi="334" he="50" img-content="drawing" img-format="GIF"/
each signal point is marked with a binary string representing a block of coded symbol values associated with the modulation symbol. In a particular embodiment, a gray code mapping is used to associate modulation symbols with blocks of code symbols. It should be appreciated that other forms of mapping may be used in place of the gray coded mapping. According to the gray-coded mapping, the labeling of the ith point on the constellation is given by:
li＝bm-1bm-2bm-3...b2b1b0wherein b isk＝0，1； img id="idf0029" file="A20071013644700272.GIF" wi="143" he="43" img-content="drawing" img-format="GIF"/
For m 1, said 2mthe-PSK constellation is a well-known set of BPSK signals. For m 2, said 2mthe-PSK constellation conforms to the 4QAM constellation shown in fig. 9 and is often referred to as the QPAK constellation. Said 2 for m-3 is depicted in fig. 12m-a PSK signal constellation. Said 2 for m-4 is depicted in fig. 13m-a PSK signal constellation.
In one embodiment, binary data at the input of the transmitter is encoded and mapped to points on a signal constellation. Thus, the modulation symbol sequence x, which can be extracted from the normalized square QAM constellation, is x1x2x3.. to advantageously model the signal at the channel input. By the sequence y ═ y1y2y3.., giving the output of the channel, whereinimg id="idf0030" file="A20071013644700273.GIF" wi="155" he="24" img-content="drawing" img-format="GIF"/
Wherein EnRepresents the average modulated signal energy at the receiver as determined by the channel gain over the nth symbol duration; thetanRepresenting a phase shift introduced by the channel over the duration of the nth symbol; and wnRepresenting a variance E * | w with a mean of 0 introduced by the channel, other users, multipath signals, etcn|2*＝Nt，nComplex additive gaussian noise (and pseudo-random (PN) interference).
Assuming that the receiver has gain on the complex channelimg id="idf0031" file="A20071013644700274.GIF" wi="91" he="24" img-content="drawing" img-format="GIF"/Sum noise variance Nt，nTo recognize that they can be estimated by transmitting pilot signals along with the data. The receiver calculates a maximum ratio dot product to produce a decision variable
img id="idf0032" file="A20071013644700275.GIF" wi="281" he="51" img-content="drawing" img-format="GIF"/
The above equation yields the following equation:
img id="idf0033" file="A20071013644700281.GIF" wi="146" he="43" img-content="drawing" img-format="GIF"/img id="idf0034" file="A20071013644700282.GIF" wi="136" he="43" img-content="drawing" img-format="GIF"/img id="idf0035" file="A20071013644700283.GIF" wi="280" he="51" img-content="drawing" img-format="GIF"/
in another way of representation, the display is,
img id="idf0036" file="A20071013644700284.GIF" wi="353" he="52" img-content="drawing" img-format="GIF"/
if several independent copies of the channel output are available at the receiver (i.e., through time (repetition), spatial (antenna), or frequency (multipath) diversity), the individual maximum ratio point products can be added together to obtain a decision variable that maximizes the detected signal-to-noise ratio (SNR). The maximum possible detection SNR of the diversity receiver is the sum of the SNRs associated with the individual maximum ratio dot products:
img id="idf0037" file="A20071013644700285.GIF" wi="193" he="48" img-content="drawing" img-format="GIF"/
img id="idf0038" file="A20071013644700286.GIF" wi="391" he="52" img-content="drawing" img-format="GIF"/
can change decision variable ZnSeen as x produced by the demodulatornSoft decision estimation.
In one embodiment, LLR calculations for square QAM constellations are performed. Each modulation symbol represents a certain string of coded symbol values defined by its label. Given that all modulation symbols are equally possible, the given code symbol b is given bykAssociated LLRs, where 0 ≦ k < 2 m:
img id="idf0039" file="A20071013644700287.GIF" wi="412" he="82" img-content="drawing" img-format="GIF"/
img id="idf0040" file="A20071013644700288.GIF" wi="199" he="82" img-content="drawing" img-format="GIF"/
img id="idf0041" file="A20071013644700289.GIF" wi="351" he="121" img-content="drawing" img-format="GIF"/
advantageously, some simplification results from the product symmetry of the squared QAM signal constellation and the gray coded marker function. In particular, it should be noted that the position c of the (i, j) th point may be seti，jWrite as ci，j＝ai+iajWherein a isi＝(2m-1-2 i). DELTA. Likewise, when k < m, the symbol b is encodedkThe value of (i, j) depends only on i, and vice versa on j. Therefore, the above expression can be simplified as follows:
img id="idf0042" file="A20071013644700291.GIF" wi="615" he="112" img-content="drawing" img-format="GIF"/
img id="idf0043" file="A20071013644700292.GIF" wi="653" he="115" img-content="drawing" img-format="GIF"/
in view of the above symmetry, it is sufficient to solve the problem of computing the code symbol bkWherein 0. ltoreq. k < m. For the 4qam (qpsk) constellation, the calculation is simplified to the well-known expression:
LLR(b0)＝4·Re[Zn]·a0＝2*·Re[Zn]，LLR(b0)＝4·Im[Zn]·b0＝2*·Im[Zn]
as will be appreciated by those of ordinary skill in the art, the LLRs associated with the various code symbols are plotted in fig. 14, 15 and 16 for the 16QAM, 64QAM and 256QAM constellations, respectively, having gray code markers, at a reasonable operating SNR.
In one embodiment, proceed with 2mLLR calculation of PSK constellations. Each modulation symbol represents a certain string of coded symbol values defined by its label. Given that all modulation symbols are equally possible, the given code symbol b is given bykAssociated LLRs, where 0 ≦ k < m:
img id="idf0044" file="A20071013644700293.GIF" wi="393" he="81" img-content="drawing" img-format="GIF"/
img id="idf0045" file="A20071013644700294.GIF" wi="181" he="81" img-content="drawing" img-format="GIF"/
img id="idf0046" file="A20071013644700295.GIF" wi="304" he="112" img-content="drawing" img-format="GIF"/
img id="idf0047" file="A20071013644700296.GIF" wi="133" he="84" img-content="drawing" img-format="GIF"/
the final simplification in the above equation is due to the fact that | c for any point on the MPSK constellationi|21. As a result, the code symbol LLRs do not depend on the normalized SNR S. If it is in the form of polar coordinatesIndicating the modulation symbol soft decisionThe above equation can be rewritten as follows:
img id="idf0049" file="A20071013644700302.GIF" wi="386" he="111" img-content="drawing" img-format="GIF"/
fig. 17 and 18 depict the code symbol LLRs for 8PSK and 16PSK constellations, respectively. In the graphs of fig. 17 and 18, the magnitude of the soft decision is kept constant, and the angle varies from 0 to 360 degrees.
In one embodiment, for a squared 0AM constellation, a piecewise linear approximation is made to the LLR associated with each code symbol. As can be seen from FIGS. 14-16, for any 4 with m > 0m-QAM constellation, function LLRm-1(x) Is an odd function of x and can be represented by the LLR at the slope and x-0m-1(x) Is approximated by a line whose slope is consistent.
Therefore, the temperature of the molten metal is controlled,
img id="idf0050" file="A20071013644700303.GIF" wi="253" he="45" img-content="drawing" img-format="GIF"/
it can be seen that
img id="idf0051" file="A20071013644700304.GIF" wi="476" he="97" img-content="drawing" img-format="GIF"/
img id="idf0052" file="A20071013644700305.GIF" wi="226" he="96" img-content="drawing" img-format="GIF"/For large S.
By substituting specific values of m, the following equation can be obtained:
img id="idf0053" file="A20071013644700306.GIF" wi="207" he="38" img-content="drawing" img-format="GIF"/if m is 1(4-QAM/QPSK)
img id="idf0054" file="A20071013644700307.GIF" wi="282" he="46" img-content="drawing" img-format="GIF"/If m is 2(16-QAM)
img id="idf0055" file="A20071013644700308.GIF" wi="300" he="47" img-content="drawing" img-format="GIF"/If m is 3(64-QAM)
img id="idf0056" file="A20071013644700309.GIF" wi="378" he="47" img-content="drawing" img-format="GIF"/If m is 4 (256-QAM).
Advantageously provides LLRk(x) Wherein k < m-1. For k 0 ≦ m-1, LLRk(x) Is an even function of x, and when x is 2k+1S Δ time, LLRk(x) 0. For small x values, the LLR can be approximated by a trigonometric functionk(x) And for large x values, a straight line approximation may be used.
img id="idf0057" file="A20071013644700311.GIF" wi="664" he="56" img-content="drawing" img-format="GIF"/
WhereinAnd
it is helpful to separate the two cases k-m-2 and k < m-2.
For k-m-2, the above expression is simplified to:
img id="idf0060" file="A20071013644700314.GIF" wi="618" he="54" img-content="drawing" img-format="GIF"/
and for k < m-2, the above expression is simplified to:
img id="idf0061" file="A20071013644700315.GIF" wi="590" he="54" img-content="drawing" img-format="GIF"/
thus, for each k, m, such that 0 ≦ k < mEnough to calculate LLRk(0)。
The case of k ═ m-2 can advantageously be handled individually:
img id="idf0062" file="A20071013644700316.GIF" wi="326" he="52" img-content="drawing" img-format="GIF"/
img id="idf0063" file="A20071013644700317.GIF" wi="344" he="53" img-content="drawing" img-format="GIF"/
for k < m-2, the following equation holds:
img id="idf0064" file="A20071013644700318.GIF" wi="451" he="75" img-content="drawing" img-format="GIF"/
img id="idf0065" file="A20071013644700319.GIF" wi="469" he="76" img-content="drawing" img-format="GIF"/
if the above result is embodied as m 2, 3 and 4, then
LLRm-2(x)|x＝0＝-8Δ2S-0.8S for m 2(16-QAM)
img id="idf0066" file="A20071013644700321.GIF" wi="348" he="62" img-content="drawing" img-format="GIF"/For m ═ 3(64-QAM)
img id="idf0067" file="A20071013644700322.GIF" wi="301" he="56" img-content="drawing" img-format="GIF"/
img id="idf0068" file="A20071013644700323.GIF" wi="314" he="63" img-content="drawing" img-format="GIF"/For m-4 (256-QAM)
img id="idf0069" file="A20071013644700324.GIF" wi="439" he="61" img-content="drawing" img-format="GIF"/For m ═ 3(64-QAM)
img id="idf0070" file="A20071013644700325.GIF" wi="292" he="54" img-content="drawing" img-format="GIF"/
img id="idf0071" file="A20071013644700326.GIF" wi="309" he="62" img-content="drawing" img-format="GIF"/For m-4 (256-QAM)
img id="idf0072" file="A20071013644700327.GIF" wi="392" he="54" img-content="drawing" img-format="GIF"/
img id="idf0073" file="A20071013644700328.GIF" wi="309" he="62" img-content="drawing" img-format="GIF"/For m-4 (256-QAM)
For large values of normalized SNR S, the above expression can be approximated as follows:
LLRm-2(x)|x＝0＝-8Δ2S-0.8S for m 2(16-QAM)
≈24Δ2S ═ (12/21) S ═ 3(64-QAM) for m
≈80Δ2S ═ (8/17) S ═ 4(256-QAM) for m
LLRm-3(x)|x＝0≈8Δ2S ═ (4/21) S ═ 3(64-QAM) for m
≈24Δ2S ═ (12/85) S ═ 4(256-QAM) for m
LLRm-4(x)|x＝0≈8Δ2S ═ (4/85) S ═ 4(256-QAM) for m
In one embodiment, for MPSK, the LLRs associated with each code symbol are triangulated. As can be seen from fig. 17 and 18, for k ═ m-1, the LLR can be approximated by a sine functionk(z) and for k < m-1 a cosine function can be used for approximation. More precisely to the fact that the pressure of the liquid,
img id="idf0074" file="A20071013644700329.GIF" wi="578" he="42" img-content="drawing" img-format="GIF"/
for k being more than or equal to 0 and less than m-1,
img id="idf0075" file="A20071013644700331.GIF" wi="484" he="25" img-content="drawing" img-format="GIF"/
img id="idf0076" file="A20071013644700332.GIF" wi="468" he="61" img-content="drawing" img-format="GIF"/
in particular, setting k-m-2 and k-m-3 will yield, respectively:
img id="idf0077" file="A20071013644700333.GIF" wi="204" he="43" img-content="drawing" img-format="GIF"/
and
img id="idf0078" file="A20071013644700334.GIF" wi="462" he="47" img-content="drawing" img-format="GIF"/
this necessarily simplifies the estimation of LLR for k < m-1m-1(jR) and LLRk(R) in the presence of a catalyst. It can be shown that:
img id="idf0079" file="A20071013644700335.GIF" wi="363" he="106" img-content="drawing" img-format="GIF"/
for k < m-2,
img id="idf0080" file="A20071013644700336.GIF" wi="557" he="77" img-content="drawing" img-format="GIF"/
the above results may be embodied as m 2, 3 and 4 to yield the following set of equations, respectively:
img id="idf0081" file="A20071013644700337.GIF" wi="298" he="81" img-content="drawing" img-format="GIF"/
img id="idf0082" file="A20071013644700338.GIF" wi="45" he="19" img-content="drawing" img-format="GIF"/for m 2(QPSK)
img id="idf0083" file="A20071013644700339.GIF" wi="201" he="81" img-content="drawing" img-format="GIF"/
img id="idf0084" file="A200710136447003310.GIF" wi="175" he="48" img-content="drawing" img-format="GIF"/For m ═ 3(8-PSK)
img id="idf0085" file="A200710136447003311.GIF" wi="380" he="81" img-content="drawing" img-format="GIF"/
img id="idf0086" file="A200710136447003312.GIF" wi="185" he="48" img-content="drawing" img-format="GIF"/For m ═ 4(16-PSK)
img id="idf0087" file="A20071013644700341.GIF" wi="272" he="81" img-content="drawing" img-format="GIF"/
img id="idf0088" file="A20071013644700342.GIF" wi="184" he="48" img-content="drawing" img-format="GIF"/For m ═ 3(8-PSK)
img id="idf0089" file="A20071013644700343.GIF" wi="362" he="81" img-content="drawing" img-format="GIF"/
img id="idf0090" file="A20071013644700344.GIF" wi="189" he="48" img-content="drawing" img-format="GIF"/For m ═ 4(16-PSK)
img id="idf0091" file="A20071013644700345.GIF" wi="272" he="81" img-content="drawing" img-format="GIF"/
img id="idf0092" file="A20071013644700346.GIF" wi="185" he="48" img-content="drawing" img-format="GIF"/For m ═ 3(8-PSK)
img id="idf0093" file="A20071013644700347.GIF" wi="362" he="80" img-content="drawing" img-format="GIF"/
img id="idf0094" file="A20071013644700348.GIF" wi="189" he="47" img-content="drawing" img-format="GIF"/For m ═ 4(16-PSK)
img id="idf0095" file="A20071013644700349.GIF" wi="446" he="80" img-content="drawing" img-format="GIF"/
img id="idf0096" file="A200710136447003410.GIF" wi="190" he="47" img-content="drawing" img-format="GIF"/For m ═ 4(16-PSK)
Thus, a novel and improved method and apparatus for computing soft decision input metrics for a turbo decoder has been described. Those of ordinary skill in the art would appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. Various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. The skilled artisan recognizes the interchangeability of hardware and software under these circumstances, and how best to implement the described functionality for each particular application. By way of example, the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, hardware components such as registers and FIFO, a processor executing a set of firmware instructions, any conventional programmable software module and a processor, or combinations thereof. The processor may advantageously be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. The software modules may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. Those of ordinary skill in the art will further appreciate that the data, instructions, commands, information, signals, bits, symbols, and symbols that may be referenced throughout the above description are advantageously represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or a combination thereof.
The preferred embodiments of the present invention have been shown and described. It will be apparent, however, to one of ordinary skill in the art that many changes can be made to the embodiments disclosed herein without departing from the spirit and scope of the invention. Accordingly, the invention is not to be restricted except in light of the following claims.

Claims (16)

1. A method of approximating log-likelihood ratio metrics for a plurality of turbo encoded symbols, the plurality of turbo encoded symbols having been modulated by square quadrature amplitude modulation signal constellations having gray code labeling, the method comprising the steps of:
extracting complex modulation symbol soft decisions on modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component;
scaling the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the most significant turbo-coded symbols of the modulation symbols; and
applying a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the remaining turbo encoded symbols of the modulation symbols.
2. The method of claim 1, wherein said scaling step comprises the steps of: the method includes dividing the plurality of encoded symbols into equally divided 1 st and 2 nd groups, scaling an in-phase component in the 1 st group, and scaling a quadrature component in the 2 nd group.
3. The method of claim 2, wherein said applying step comprises the steps of: applying a linear combination of a trigonometric function and a ramp function to the in-phase component of the group 1 and applying a linear combination of a trigonometric function and a ramp function to the quadrature component of the group 2.
4. The method of claim 1, wherein said applying step comprises the steps of: the method includes dividing the plurality of turbo encoded symbols into equally divided 1 st and 2 nd groups, applying a linear combination of a trigonometric function and a ramp function to an in-phase component of the 1 st group, and applying a linear combination of a trigonometric function and a ramp function to a quadrature component of the 2 nd group.
5. The method of claim 4 wherein said scaling step comprises the steps of scaling the in-phase component in said 1 st group and scaling the quadrature component in said 2 nd group.
6. A receiver configured to approximate log-likelihood ratio metrics for a plurality of turbo encoded symbols, the plurality of turbo encoded symbols having been modulated by square quadrature amplitude modulation signal constellations having gray code labeling, the receiver comprising:
means for extracting complex modulation symbol soft decisions on received modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component;
means for scaling the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the most significant turbo-coded symbols of the modulation symbols; and
means for applying a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the remaining turbo-encoded symbols of the modulation symbols.
7. The receiver of claim 6, wherein:
the means for extracting is a demodulator; and
the means for scaling and applying is a log likelihood ratio calculation module.
8. The receiver of claim 7, wherein the log-likelihood ratio calculation module is further configured to scale the complex modulation symbol soft decisions by dividing the plurality of turbo encoded symbols into equally divided 1 st and 2 nd groups, scaling an in-phase component in the 1 st group, and scaling a quadrature component in the 2 nd group.
9. The receiver of claim 8 wherein the log-likelihood ratio computation module is further configured to apply a linear combination of trigonometric and ramp functions to the complex modulation symbol soft decisions by applying a linear combination of trigonometric and ramp functions to the in-phase components of the 1 st group and applying a linear combination of trigonometric and ramp functions to the quadrature components of the 2 nd group.
10. The receiver of claim 7, wherein the log-likelihood ratio calculation module is further configured to apply a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions by dividing the plurality of turbo encoded symbols into equally divided 1 st and 2 nd groups, applying a linear combination of a trigonometric function and a ramp function to the in-phase component of the 1 st group, and applying a linear combination of a trigonometric function and a ramp function to the quadrature component of the 2 nd group.
11. The receiver of claim 10 wherein the log-likelihood ratio computation module is further configured to scale the complex modulation symbol soft decisions by scaling the in-phase components in the 1 st group and scaling the quadrature components in the 2 nd group.
12. A receiver configured to approximate log-likelihood ratio metrics for a plurality of turbo encoded symbols, the plurality of turbo encoded symbols having been modulated by square quadrature amplitude modulation signal constellations having gray code labeling, the receiver comprising:
a processor; and
a processor-readable storage medium coupled to the processor and containing a set of instructions executable by the processor for extracting complex modulation symbol soft decisions on received modulation symbols, the modulation symbols being associated with a plurality of turbo encoded symbols, the complex modulation symbol soft decisions having an in-phase component and a quadrature component; scaling the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the most significant turbo-coded symbols of the modulation symbols; and applying a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions to obtain log-likelihood ratio metrics for the remaining turbo encoded symbols of the modulation symbols.
13. The receiver of claim 12, wherein the set of instructions are further executable by the processor to scale the complex modulation symbol soft decisions by partitioning the plurality of turbo encoded symbols into equally partitioned 1 st and 2 nd groups, scaling an in-phase component in the 1 st group, and scaling a quadrature component in the 2 nd group.
14. The receiver of claim 13, wherein the set of instructions are further executable by the processor to apply a linear combination of trigonometric and ramp functions to the complex modulation symbol soft decisions by applying a linear combination of trigonometric and ramp functions to the in-phase component of the group 1 and applying a linear combination of trigonometric and ramp functions to the quadrature component of the group 2.
15. The receiver of claim 12, wherein the set of instructions are further executable by the processor to apply a linear combination of a trigonometric function and a ramp function to the complex modulation symbol soft decisions by dividing the plurality of turbo encoded symbols into equally divided 1 st and 2 nd groups, applying a linear combination of a trigonometric function and a ramp function to an in-phase component of the 1 st group, and applying a linear combination of a trigonometric function and a ramp function to a quadrature component of the 2 nd group.
16. The receiver of claim 12, wherein the set of instructions are further executable by the processor to scale the complex modulation symbol soft decisions by scaling an in-phase component in the set 1 and scaling a quadrature component in the set 2.
HK08102436.0A 2000-03-08 2008-03-04 Method and apparatus for computing soft decision input metrics to a turbo decoder HK1111835A (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US09/521,358 2000-03-08

Publications (1)

Publication Number Publication Date
HK1111835A true HK1111835A (en) 2008-08-15

Family

ID=

Similar Documents

Publication Publication Date Title
JP4741154B2 (en) Apparatus and method for calculating a metric of a soft decision input for a turbo decoder
US6377607B1 (en) System and method for performing accurate demodulation of turbo-encoded signals via pilot assisted coherent demodulation
CN1048837C (en) Method and decoder circuit and system for decoding non-coherently demodulated signals
US6977977B1 (en) Compensation of I/Q gain mismatch in a communications receiver
US6944242B2 (en) Apparatus for and method of converting soft symbol information to soft bit information
EP1350310B1 (en) Reduced soft output information packet selection
EP1946509B1 (en) Method and apparatus for normalizing input metric to a channel decoder in a wireless communication system
EP1195908A2 (en) Method and apparatus for processing modulation symbols for soft input decoders
WO2009047755A2 (en) Decoding of forward error correction codes in the presence of phase noise
GB2395097A (en) A decoder apparatus and method of decoding therefor
CA2434806A1 (en) Apparatus and method for symbol combining in a mobile communication system
US6901121B1 (en) Compensation of DC offset impairment in a communications receiver
US20050129147A1 (en) Method and system for modulating and detecting high datarate symbol communications
EP1198938A1 (en) Fading resistant multi-level qam receiver
CN101595700B (en) Log-likelihood ratio (llr) computation using piecewise linear approximation of llr functions
US5608763A (en) Method and apparatus for decoding a radio frequency signal containing a sequence of phase values
WO2011098297A1 (en) Process for performing log-likelihood-ratio clipping in a soft-decision near-ml detector, and detector for doing the same
Valenti et al. Iterative multisymbol noncoherent reception of coded CPFSK
US8422600B2 (en) Apparatus and method for estimating phase error based on variable step size
JP2007166592A (en) Soft decision demapping method suitable for high-order modulation scheme for iterative decoder and error correction apparatus using the same
HK1111835A (en) Method and apparatus for computing soft decision input metrics to a turbo decoder
JP4139706B2 (en) Receiver
Nam et al. LDPC codes with BDPSK and differential detection over flat Rayleigh fading channels
KR20070050750A (en) Method and device for metric normalization input to channel decoder in wireless communication system
US20030112899A1 (en) Decision directed phase locked loops (DD-PLL) with excess processing power in digital communication systems