JPH06224878A - Code division multiplex communication method and code division multiplex communication receiver - Google Patents

Abstract

(57) [Summary] (Modified) [Purpose] To increase the number of N in 1-to-N communication. [Structure] Each terminal station 12 ₁ divides a binary information sequence into frames, each frame is error-correction coded, and a 2 ^M- value random sequence unique to each terminal station is added to it by an adder circuit 22. One of the 2 ^M spreading sequences is transmitted from the transmitter 23 for each symbol. Matched filter 26 of central station 11
Despreading each spreading sequence in 1~26-2 ^M, subtraction of 2 ^M random sequence of each terminal station number Yes from the value compared with a threshold value to each of its output by the comparator 29-1～29-2 ^M Circuit 31 ₁ ~
31 _N is subtracted, and the decoders 32 _{1 to} 32 _N perform Viterbi decoding to obtain m decoded codeword candidates. A plurality of combinations in which one candidate is selected from each decoder are created by the selector 33, the reception pattern candidates and the output patterns of the comparators 29-1 to 29-2 ^M are compared by the circuit 35, and the number of mismatched symbols is 1 frame. Then, each decoded codeword corresponding to the reception pattern candidate having the minimum value is correctly output.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention provides 1: N multiplex communication.
The present invention relates to a code division multiplex communication method performed by code division spread spectrum communication, and a code division multiplex communication receiver in its central station.

[0002]

2. Description of the Related Art A code division multiplex communication system using spread spectrum communication is applied to various communication systems.
Or is being done. Especially in mobile communication systems, the code division multiplex (CDMA) system can significantly improve the spectrum utilization rate as compared with the systems based on the frequency division multiplex (FDMA) system and the time multiplex (TDMA) system currently in operation. That there is a possibility of reference 1: KS Gilho
usen, IM Jacobs, R.Padovani, A.Viterbi, LAWea
ver, Jr. and CEWheatley III, "On the Capacity of a
Cellular CDMA System ", IEEE Trans. VT, Vol. VT-40,
pp.303-312,1991, the examination for practical use is becoming more and more vigorous.

By the way, generally, increasing the spectrum utilization rate of a system is nothing but increasing the number of simultaneous telecommunications users (corresponding to N above, hereinafter referred to as capacity) permitted by the system. C shown in the above literature 1 etc.
The basic design concept of the DMA system is that signals from other users are regarded as noise. Even with this assumption, the estimated system capacity of CDMA is FDMA.
And TDMA.

Now, the capacity of this CDMA system can be dramatically increased by not considering the signals from other users as noise, but rather by effectively using its knowledge.
That is, a filter that minimizes the influence of interference from other users can be configured by using the correlation between the signal from another user and the desired signal of the user (for example,
Reference 2: R. Lupas and S. Verdu, "Near-Far Resistac
e of Multiuser Detectors in Asynchronous Channel
s ", IEEE Trans. COM, Vol. COM-38, pp.496-508, 1990
reference). In this way, multiple communication environments (such as FDMA and TD
Instead of using the MA method or a stack of independent channels (which operate independently of other channels) as in the above-mentioned paper of KSGilhousen et al. Communication methods that share are known as "multi-user channels." A receiver that decodes signals of a plurality of simultaneous users by using the characteristics of such a multi-user channel is called a multi-user receiver.

Now, there are two methods for realizing a system that specifically uses the characteristics of such a multi-user channel. One uses channel additivity. That is, a signal transmitted from a plurality of simultaneous users is expressed as a sum of these combined signals at the receiving station (if the received signal is expressed as a complex envelope, the combined signal is expressed as the sum of complex signals. Is additive). By utilizing this, by removing the estimated waveform of the received signal of the user of interest from the received signals which are the combined waveforms of all the users, the influence of the received signal of this user on the combined signal can be removed. The received signal of the user of interest is obtained by convolving the estimated transmitted signal of the user obtained by despreading with the estimated value of the transfer function of the channel. This method is conventionally known as a replica reproduction type interference cancellation. The above papers by R. Lupas and S. Verdu can also be said to be methods based on this property of the channel.

By the way, when a channel for communication is composed of a plurality of propagation paths as in a mobile communication environment,
As described above, configuring the multi-user receiver based on the additivity of the channel has a limit in the achievable performance in that the required processing becomes enormous. Another method of realizing a system that specifically uses the characteristics of a multi-user channel is a method of giving a channel a characteristic of OR (logical sum) logic. In a channel having an OR logical characteristic (hereinafter referred to as an OR channel), information transmitted via the channel is represented by a binary value of 0 or 1, and information of one or more users among a plurality of simultaneous users is represented. Has a characteristic that the channel output becomes 1. Therefore, from the viewpoint of the user who has transmitted 1, 1 is not affected even if another user similarly transmits 1. By using such characteristics of the channel and an error correction code suitable for this, it is possible to reproduce the signal for each user from the sequence after the received signal is binary-identified (for example, Reference 3: Kawahara, Matsumoto). "DS / CDMA mobile radio system with OR channel characteristics", Proceedings of 1992 Autumn Meeting of the Institute of Electronics, Information and Communication Engineers). This method is applicable not only in a multipath environment in which a channel is composed of a plurality of propagation paths, but also in the near-far problem in a CDMA system (the reception power of a signal from a terminal station near the central station is
A signal from a distant terminal station cannot be received because it is much larger than the received power of the signal from a distant terminal station. In order to avoid this problem, highly accurate transmission power control is needed. The feature is that the accuracy required for the control error of the transmission power control for solving the above problem is greatly relaxed.

This will be described in more detail below.
Now, 2 ^M spreading sequences are prepared for a channel, and a plurality of terminal stations select one of them according to a transmission signal and transmit it. Therefore, the channel has the ability to separately transmit M bits / symbol. In the reception processing at the central station, first, despreading is performed using a matched filter that matches each of these spread sequences received. The output of the matched filter is subjected to energy detection (identifying whether the received power of each spreading sequence is above or below a predetermined threshold). From the viewpoint of a terminal station that has transmitted a certain spreading sequence, even if another terminal station transmits the same spreading sequence, energy is still detected in the despreading output for this spreading sequence. It has been configured.

An error correction code having an efficiency of 1 / M is used in order to detect the transmission signal from the desired terminal station from the output of the OR channel. Therefore, one symbol is an M-bit error correction code. At the central station, the error correction code is decoded for the above 2 ^M energy detection outputs (maximum likelihood decoding by the hard-decision Viterbi algorithm is performed),
In the decoding process, the entry to the non-transmission path on the trellis for the code of the desired terminal station is assumed to occur randomly. That is, specifically, randomization is performed by adding a random sequence of 2 ^M values different from each other to the output sequence of the error correction encoder with an efficiency of 1 / M at each terminal station, modulo 2 ^M. Has been achieved. Therefore, the transmission patterns (transmission codes) of other terminal stations are not considered at all, and the correlation between the energy detection output for the 2 ^M spreading sequences and the transmission pattern of the entire simultaneous transmission terminal station also affects the decoding result. Absent.

[0009]

By providing the channel with the OR logical characteristic as described above, there is an advantage that a multi-user receiver can be easily configured even in a multi-path propagation environment. However, in the conventionally known configuration method of the multi-user receiver, such as the method of Document 3,
The total bit rate that a channel can transmit (called Rate Sum) has a drawback that it does not become so large as compared with the method of stacking individual independent channels to form a multiple communication environment. For example, the increase in bit rate by the multi-user receiver shown in the above Document 3 is only 35%. In other words, the conventional method does not fully utilize the characteristics of the channel having the OR logic characteristic.

[0010]

According to the invention of claim 1, the central station transmits the reference symbol timing and the frame timing having a period of a plurality of symbols to N (N is an integer of 2 or more) terminal stations. Then, in each terminal station, the transmission information sequence is divided into frames each having a length equal to the reception frame timing period to form a code word, and an efficiency of 1 is obtained for each code word.
/ Encoding of M (M is an integer of 2 or more) is performed, to the coding sequence, the same speed and this, by adding the scrambling random sequence of 2 ^M values of Koyu to each terminal station of 2 ^M modulo hand,
A transmission code string selected from 2 ^M transmission codes is used, and this transmission code string is synchronized with O
Transmit to the central station via R channel.

[0011] In the central station from the received signal by 2 the ^M determiner, 2 each one of the ^M transmission code is received whether each determined for each symbol, received by those 2 the ^M determiner for those determined to have, 2 ^M of each individual chromatic of each terminal station from the value of 2 ^M values given to the determination unit
A scrambled random sequence of values is subtracted modulo 2 ^M , and maximum likelihood decoding is performed on the N subtraction results for a code with an efficiency of 1 / M. Decoding codeword candidates are determined, for example, decoding codewords whose likelihood is greater than or equal to a predetermined value are taken as candidates, and these candidates are taken out one by one from N decoding systems, and their corresponding transmission codes are created to obtain these. The reception pattern candidates are combined to create reception pattern candidates for different combinations of codeword candidates, and these reception pattern candidates are compared with the actual N combined decoded patterns for each symbol timing. Then, a reception pattern candidate with the smallest total sum of the number of different symbols over one frame is obtained, and N decoding results corresponding to this are received, and the transmission signals from the corresponding N terminal stations are received. And decoding result.

According to the invention of claim 2, the method of determining a correct reception pattern is different from that of the invention of claim 1, and the soft decision values of 2 ^M decision elements for each symbol timing which are not included in each reception pattern candidate. Is calculated over one frame, and the reception pattern candidate having the minimum total is set as a correct reception pattern. According to the invention of claim 3, claim 1
Or in the invention of 2, the terminal station is encoded with a constraint length K.
(K is an integer of 2 or more) is encoded into a convolutional code, and the transmission information sequence is divided from the frame timing cycle by (K-
1) The number of symbols is reduced, and the state transition of the convolutional code is converged to a predetermined state on the trellis continuously in each divided transmission information sequence. A tail symbol of (K-1) symbols is transmitted to the central station. The maximum likelihood decoding is performed on the trellis that converges to the predetermined state at the symbol timing corresponding to the tail symbol of (K-1) symbols.

According to a fourth aspect of the present invention, in the first or second aspect of the invention, each terminal station divides the transmission information sequence into a length that is L (L is an integer of 1 or more) symbols less than the frame timing period, and the division is performed. An error detection check symbol having a length of L symbols is added to each transmission information sequence, and the central station performs error detection on a plurality of probable codeword candidates obtained in each decoding process. Reception pattern candidates are generated from word candidates.

[0014]

When the 2 ^M spreading sequence described in the section of the prior art is used as the OR channel, transmission patterns from all terminal stations (combined pattern of transmission codes transmitted from all terminal stations) are used in the present invention. Then, decoding is performed in consideration of the correlation with the energy detection output for the 2 ^M spreading sequences.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described with reference to FIG.
It The central station 11 has N terminal stations 12 ₁~ 12_N(In the figure
12₁1-to-N phase-directional communication with
ing. Decoding at the central station 11 consists of multiple symbols in one cycle.
The same frame timing for N terminal stations.
The symbol timing is synchronized in all terminal stations.
It is assumed that they are gently synchronized. So in this example
The central station 11 uses the downlink channel transmitter 13,
A symbol indicating the frame timing together with the downlink information signal
All terminal stations 12₁~ 12_NDown channel 14
To send over. Terminal station 12₁~ 12_NThen down channel
Signal from the central station 11 via the channel 14 for downlink channel
It is received by the receiver 15 to reproduce the symbol timing and
The frame timing is reproduced and supplied to the transmission unit 16,
Is used for the timing of upstream signal transmission, and the beginning of the frame is
The indicated symbol timing is transmitted. Actually, the propagation path
Due to the difference in the delay time of the
Jitter occurs, but it is all over within the time range of the propagation time difference.
Signal can be received.

Each of the terminal stations 12 _{1 to} 12 _N transmits information frame by frame in synchronization with the reproduction frame timing. That is, the frame assembling circuit 17 in the transmitter 16 synchronizes the binary information sequence from the terminal 18 with the symbol at the reproduction symbol timing and divides (framing) a plurality of symbols at the reproduction frame timing as a cycle. . Each divided information sequence is encoded by an encoding circuit 19 into an error correction code having an efficiency of 1 / M. The output of the encoding circuit 19 is an M-bit code for each symbol, that is, one of 2 ^M values.
From the 2 ^M- value randomized sequence generation circuit 21 to each terminal station 12 ₁
2 ^M for each unique scramble at ~ 12 _N
A value random sequence is generated in synchronism with the reproduction symbol timing, and this 2 ^M- value random sequence is added to the 2 ^M- value sequence output from the encoding circuit 19 by adding 2 modulo 2 ^M. Done in. That is, in each terminal station, the coded value of each symbol is shifted by 0 to 2 ^M values, the shifted value is changed randomly, and the random change is different for each terminal station.

2 ^M for each symbol of the adder circuit 22
The transmission code of any one of the values is transmitted to the central station 11 via the OR channel. In this example, 2 as the OR channel
Output 2 of adder circuit 22 when ^M spreading sequences are used
^{Using the M} value as an address, one of 2 ^M spreading sequences in the transmitter 23 is selected for each symbol and transmitted to the central station 11. It is assumed that the central station 11 knows the generated random sequence of the 2 ^M- value randomized sequence generation circuit 21 of each of the terminal stations 12 _{1 to} 12 _N.

At the central station 11, each terminal station 12₁~ 12_NOr
From the transmitted upstream channel spreading sequence 24 to the receiver 25
Received in 2^M2 matched to each of the spreading sequences^M
Number of matched filters 26-1 to 26-2^MSupply to each expansion
Performs despreading on the scattered sequence. These despread outputs 2
7-1 to 27-2^MAnd terminals 28-1 to 28-2^MFrom
Compared to each threshold (these thresholds are usually the same value)
Vessel 29-1 to 29-2 ^MIs greater than the threshold value
Outputs 1 from the comparator and 0 from the comparator when it is small.
It Matched filters 26-1 to 26-2^MIs given to self
When the spread sequence is input, the despread output has large power.
Become. Each set of these matched filters and comparators is claimed.
Corresponds to the determiner in the inventions 1, 5; that is, 2^MOf
Each judge is 2^MIndividual transmission codes (transmitter 23 input
Of the received signal and associated with any one of
When the transmission code given to itself is detected, 1 is output.

2 ^M- value randomized sequence generation circuits 30 _{1 to} 3
From 0 _N , the same 2 ^M value random sequence as the generated 2 ^M value random sequence of the 2 ^M value randomized sequence generation circuit 21 of each of the terminal stations 12 _{1 to} 12 _N is synchronously generated, and these N
The 2 ^M- value random sequences are compared by comparators 29-1 to 29-2.
^{For one of the M} outputs, the comparator 29-1 to 29-1
2 ^M from any of the 2 ^M values given to 29-2 ^M
The subtraction modulo is performed at each symbol timing. As a result, a signal corresponding to the 2 ^M- value coded signal at the input side of the adder circuit 22 of each of the terminal stations 12 _{1 to} 12 _N can be obtained. Each of the output subtraction sequences of these subtraction circuits 31 _{1 to} 31 _N is independently decoded by the decoders 32 _{1 to} 32 _N by the Viterbi algorithm of the error correction code with an efficiency of 1 / M, and is passed at the end of the decoding of one frame. Metrics, that is, m codewords with high likelihood are output as candidates. Each decoder may output a code word having a higher likelihood than a predetermined likelihood, and in that case, if the number is too large, it may be limited to a predetermined value or less.

In the combination selector 33, a plurality of combinations in which code word candidates are selected one by one from the decoders 32 _{1 to} 32 _N are created and supplied to the replica generation circuit 34. For each combination of the code word candidates, the replica generation circuit 34 generates these code word candidates corresponding to the output transmission code of the adder circuit 22 on the assumption that these code word candidates are transmitted by the corresponding terminal station, and the generated transmission code is generated. Inside 2
And 1 for the corresponding transmission code if there is a match with the ^M transmission symbols, and 0 if not, the 1 or 0 2 ^M
A reception pattern candidate in which the transmission codes are arranged in order is generated. Each of the reception pattern candidates and the comparators 29-1 to 29-
The 2 ^M output pattern (actual reception pattern) is compared by the pattern comparison circuit 35 for each symbol, and the mismatched symbols (there is no symbol entry for which energy is to be detected and the energy is The number of non-detected symbols has an entry), the reception pattern candidate having the smallest total number of mismatch patterns (the number of different symbols) is selected, and the decoders 32 _{1 to} 32 _N corresponding to the reception pattern candidates are selected. Each of the decoded codewords is output as a received decoded word for the transmission signals from the terminal stations 12 _{1 to} 12 _N.

That is, focusing on a certain terminal station, if the value of the randomized sequence at each symbol timing is subtracted modulo 2 ^M from the energy detection output pattern for 2 ^M spread sequences, on the trellis of the error correction code, There is an entry with a correlation of 1 (energy is detected) for the path of the codeword transmitted by this terminal station, whereas an entry on the non-transmission path occurs at random (other terminal stations). It depends on the entry from). In the absence of noise or fading, the transmission path will have the maximum path metric (likelihood) at the end of one frame.
On the other hand, if there are entries from other terminal stations in all the symbols from the branch to the confluence of the trellis on one or more non-transmission paths, then these non-transmission paths also have the maximum path metric at the same time. Become. Further, if there is noise or fading, there may be a symbol (sequence) whose energy is not detected on the transmission path, or conversely, energy may be detected for a symbol that is not transmitted by any terminal station.
In such a case, the transmission path does not always have the maximum path metric at the end of one frame.

However, in the present invention, a plurality of code words having a high path metric (likelihood) are selected as candidates from the code words obtained as a result of Viterbi decoding for each terminal station as described above (as described above). Thus, due to the randomized sequence, there is an entry with a correlation of 1 for the transmission path, and an entry on the non-transmission path occurs randomly.
2 ^M received from all or some of these combinations).
The combination with the highest probability of generating the energy detection output pattern of each series is selected to be the final decoding result.

In the above description, reception pattern candidates are not necessarily created for all combinations of code word candidates, or m ^N combinations in the case of m combinations, for example, groups of similar ones in code word candidates of the same decoder are grouped. Then, a reception pattern candidate is created by selecting one representative from each group, and when narrowing down the reception pattern candidates, the group having a large number of different symbols is removed to reduce the number of reception pattern candidates and reduce the processing time. It can be shortened.

In the above description, the reception pattern candidate having the smallest total sum of the number of different symbols between the reception pattern candidate and the actual reception pattern is obtained, but the matched filters 26-1 to 26-
It is also possible to obtain, for each symbol, the one corresponding to 0 in the reception pattern candidates in each output power of 2 ^M , and obtain the smallest reception pattern candidate in the total sum at the end of one frame of that power. This is an embodiment of the invention of claims 2 and 6, and each output of the matched filters 26-1 to 26-2 ^M corresponds to the soft decision output of each decision unit.

In the present invention, the decoding process is 1
It is executed on a frame-by-frame basis, and an error correction code with an efficiency of 1 / M is used to select a candidate for a transmission path, and this error correction code is subjected to maximum likelihood decoding using a Viterbi algorithm on the receiving side. It is desirable to use a convolutional code as the code. However, since a convolutional code usually has an infinite memory (the current transmission symbol is affected by the past infinite time in principle), the path with high likelihood at the end of one frame is not always close to the transmission path. Not necessarily (it occurs because one frame that originally had the possibility of being reversed after that was ended in one frame). In order to avoid this problem when the convolutional code is burst-transmitted in units of one frame, a method of transmitting tail bits is known. That is, a symbol (called a tail symbol) for converging the state transition of the convolutional code to a predetermined state (normally 0 state) is transmitted on the transmission side in succession to the transmission information sequence. The length of the tail symbol is equal to (K-1) symbols for the constraint length K of the convolutional code. Also in the present invention, this problem is avoided by transmitting tail symbols. This is the invention of claim 3.

FIG. 2 shows a frame structure when transmitting tail symbols. The pre-encoding sequence 41 is a pre-encoding symbol (normally 0) for converging the state transition to a predetermined state with respect to the binary transmission information sequence INF before encoding.
Add ZERO. This symbol ZERO has a (K-1) symbol length when the constraint length of error correction coding is K symbols. In the encoded sequence 42, the tail symbol TS follows the encoded 2 ^M- value transmission information sequence CODED-INF.

For example, when m candidates having a high path metric are selected for all terminal stations and all of these combinations are compared, the number of reception pattern candidates to be created is m ^N , and the processing amount is remarkably large. May be. Therefore, according to the invention of claim 4, encoding is performed by adding a code for error detection (so-called an inner code such as a parity bit) to each of the divided transmission information sequences in advance, and the reception side has a high path metric m. For all or some of these codeword candidates, first decode the inner code, that is, check whether there is an error, and select only the codeword candidates that have not detected an error. It is used to create reception pattern candidates, and the number of reception pattern candidates is greatly reduced.

Codes for performing inner coding for error detection
The transmission information sequence before encoding is shown in FIG. Binary before encoding
Checking the transmission information sequence INF for error detection
Part CRC is added. In the above, the OR channel is 2
^MAlthough it is composed of one spreading sequence, another OR channel is used.
You can also do it. Eg 2^MYou can use individual frequencies
Yes. That is, 2 from the adder circuit 22^MDepending on the transmitted code of the value
2^MMatch any one of the frequencies
Filter 26-1 to 26-2^MInstead of above 2 each^MIndividual
Select the frequency of 2^MBandpass filters and their
Square detection of each output 2 ^MWith one detector
Just do it.

[0029]

As described above, according to the present invention, a plurality of probable codeword candidates obtained in the decoding process are extracted from the decoder corresponding to each terminal station, and one is selected from each of them. For each combination, the reception pattern candidates are created by combining the transmission codes when these are transmitted, and the reception pattern candidate that most matches the actual reception pattern is obtained, and each decoded codeword corresponding to this is obtained. Since it is correct, it is possible to expect a significant increase in system capacity.

The present invention utilizes the OR logic of the channel, as in the system of Document 3. Therefore, it has characteristics that it can be applied even in a multipath environment in which a channel is composed of a plurality of propagation paths, and that the accuracy required for transmission power control of a CDMA system is greatly relaxed. FIG. 4 shows the result of a simulation performed to show the effectiveness of the present invention. The number of sequences 2 ^M is 64, and an error correction code having a constraint length of 7 and an efficiency of 1 / M = 1/6 is used. The horizontal axis represents the number of terminal stations N, and the vertical axis represents the bit error rate after decoding. In the figure, CONV is the method shown in Reference 3, and NEW is the method according to the present invention when an error detection code is used. For example, comparing the number of users (terminal stations) that achieve a bit error rate of 10 ⁻³ , the method of Document 3 can accommodate about 25 stations, while the method of the present invention can accommodate up to about 45 stations.

[Brief description of drawings]

FIG. 1 is a diagram showing an embodiment of each invention of claims 1 and 5;

FIG. 2 is a diagram showing a frame structure when a tail symbol is transmitted.

FIG. 3 is a diagram showing a transmission information sequence before encoding when inner encoding is performed for error detection.

FIG. 4 is a diagram showing a simulation result for showing the effectiveness of the present invention.

Claims (6)

[Claims] 1. A central station transmits a reference symbol timing and a frame timing having a period of a plurality of symbols to N (N is an integer of 2 or more) terminal stations, and each of these N terminal stations is transmitted. Divides the transmission information sequence into frames having a length equal to the received frame timing cycle to obtain code words, and encodes each code word into a code of efficiency 1 / M (M is an integer of 2 or more), its coding sequence, the same speed and this, of 2 ^M scrambling random sequence of 2 ^M values of Koyu to each terminal station performs addition modulo, 2 ^M pieces of received code selected from transmission code This transmission code sequence is transmitted to the central station through a channel having a logical sum characteristic at the received reference symbol timing, and the central station transmits these channels with these 2 ^M decision devices. From the received signal The 2 ^M-number of transmission code each one of determining whether it receives each of these 2 for ^M decision unit that is determined to have been received, the value of 2 ^M values given to the determination unit From each of the above-mentioned individual scrambling random sequences of 2 ^M values of each terminal station by modulo 2 ^M , and the maximum likelihood decoding for the code of efficiency 1 / M is performed on each of the N subtraction results. , Probable decoded codeword candidates obtained in each decoding process are taken out one by one, the transmission codes corresponding to them are respectively created, and these are combined to make a reception pattern candidate. Each received pattern candidate having a different combination of the above-mentioned decoded codeword candidates and the synthesized pattern obtained by the above decoding are compared for each symbol timing, and the one frame Sum of differences symbol number determining the minimum of the received pattern candidate that which the corresponding each of N
A code division multiplex communication method, in which each of the decoding results is the reception decoding result of the transmission signals from the corresponding N terminal stations. 2. A central station transmits a reference symbol timing and a frame timing having a period of a plurality of symbols to N (N is an integer of 2 or more) terminal stations, and each of these N terminal stations is transmitted. Are coded by dividing the transmission information sequence into frames having a length equal to the received frame timing cycle, coding each codeword into a code of efficiency 1 / M (M is an integer of 2 or more), and the coding sequence, the same speed as this, the selected code sequence from 2 ^M values of the scramble random sequence of 2 ^M pieces performed addition modulo of 2 ^M transmission code Koyu to each terminal station, This transmission code string is transmitted to the central station via a channel having a logical sum characteristic in synchronism with the received reference symbol timing, and the central station receives via these channels with 2 ^M decision elements. From the signal, above 2 ^M It is determined whether or not each one of the transmission codes has been received, and for those that are determined to have been received by these 2 ^M determiners, from the value of the 2 ^M value given to that determiner, A unique scrambled random sequence of 2 ^M values of each terminal station is subtracted modulo 2 ^M , and maximum likelihood decoding is performed on the N subtraction results for a code with an efficiency of 1 / M. A plurality of probable decoded codeword candidates obtained in the decoding process are taken out one by one, the transmission codes corresponding to each are created, and these are combined to create a reception pattern candidate. A reception pattern candidate that does not exist in each reception pattern candidate having a different combination of decoded codeword candidates and has the smallest total sum of the soft decision values of the above-mentioned decision unit for one frame over one frame is obtained, and corresponding to this. That the respective N number of decoding result, and receiving the decoding result of the transmission signal from the corresponding N-number of terminal stations, code division multiple access communication method. 3. The coding of each terminal station is a convolutional code with a constraint length K (K is an integer of 2 or more), and the division of the transmission information sequence is (K-1) from the frame timing period. The number of symbols is reduced, and the tail symbols of (K-1) symbols for transmitting the state transition of the convolutional code to a predetermined state on the trellis are continuously transmitted to each of the divided transmission information sequences, and the central station transmits the tail symbols. Is a maximum likelihood decoding for a convolutional code with a constraint length K having an efficiency of 1 / M, and the maximum likelihood decoding is performed on the trellis that converges to the predetermined state at the symbol timing corresponding to the tail symbol of (K-1) symbols. 3. The code division multiplex communication method according to claim 1, wherein the method is executed. 4. Each terminal station divides the transmission information sequence into a length that is L (L is an integer greater than or equal to 1) symbols less than the frame timing period, and each divided transmission information sequence has a length of L symbols. An error detection check symbol is added, and the central station performs error detection on a plurality of probable codeword candidates obtained in each decoding process, and selects the reception pattern candidate from the codeword candidates in which no error is detected. The code division multiplex communication method according to claim 1 or 2, wherein the generation is performed. 5. A transmission information sequence is divided into frames to form codewords, each codeword is encoded into a code having an efficiency of 1 / M (M is an integer of 2 or more), and the coded sequence has 2 ^M values. the code sequence selected by performing the addition modulo of 2 ^M random sequence from 2 ^M number of transmitting codes, the code division multiplexing receiving apparatus for receiving over a channel with a logical OR characteristics, from the received signal, respectively The 2 ^M decision units for determining whether or not one of the given 2 ^M transmission codes is received, and the decision unit of the 2 ^M decision units determined to be received is given to the decision unit. the value of 2 ^M values that are, a subtractor of N for performing subtraction modulo of 2 ^M random sequence of 2 ^M values of Koyu to each terminal station (N is an integer of 2 or more), subtractors N decodings that respectively perform maximum likelihood decoding for a code with an efficiency of 1 / M for the subtraction result of Encoder, means for selecting one or more likely decoded codeword candidates obtained in the decoding process of each of these decoders, and each codeword candidate selected from these N decoders one by one. Means for taking out and creating the transmission code for each of the respective code word candidates and synthesizing these to generate a reception pattern candidate for different combinations of the code word candidates; A means for comparing each of the pattern candidates and the synthetic pattern obtained by the actual decoding at each symbol timing to obtain the total sum of the number of different symbols over one frame, and the reception pattern candidate having the smallest total sum of the number of different symbols. And means for making the respective decoding results of the N decoders corresponding to the above as the reception decoding results of the transmission signals from the N terminal units, respectively. Code division multiple access communication receiver for Bei. 6. A transmission information sequence is divided into frames to form codewords, each codeword is encoded into a code having an efficiency of 1 / M (M is an integer of 2 or more), and ^2M- valued codes are included in the encoded sequence. the code sequence selected by performing the addition modulo of 2 ^M random sequence from 2 ^M number of transmitting codes, the code division multiplexing receiving apparatus for receiving over a channel with a logical OR characteristics, from the received signal, respectively The 2 ^M decision units for determining whether or not one of the given 2 ^M transmission codes is received, and the decision unit of the 2 ^M decision units determined to be received is given to the decision unit. the value of 2 ^M values that are, a subtractor of N for performing subtraction modulo of 2 ^M random sequence of 2 ^M values of Koyu to each terminal station (N is an integer of 2 or more), subtractors N decodings that respectively perform maximum likelihood decoding for a code with an efficiency of 1 / M for the subtraction result of And a means for selecting one or more likely decoded codeword candidates obtained in the decoding process of each of these decoders, and one for each codeword candidate selected from these N decoders. , A means for performing the above-mentioned transmission code for each of the respective code word candidates and synthesizing these to generate a reception pattern candidate for different combinations of the above code word candidates; For each of the candidates, means for obtaining the total sum of the soft decision values of the decision unit for each symbol timing over one frame, and the N decodings corresponding to the reception pattern candidates having the smallest total soft decision value And a means for making each decoding result of the receiver a reception decoding result of the transmission signals from the N terminal stations.