JP3201962B2 - Variable data rate communication device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a variable data rate communication device capable of performing communication at a data rate selected for each frame among a plurality of predetermined data rates that can be transmitted.

[0002]

2. Description of the Related Art Conventionally, in order to perform variable data rate transmission, generally, data rate information of data to be transmitted is notified to a receiving side by some method. Also, in the field of mobile communication, a transmission method that does not notify the data rate of data to be transmitted to a receiving side when performing variable data rate transmission is known. For example, IS-95 standardized in the United States as a CDMA cellular telephone system using the spread spectrum of the DS (Direct Sequence) system in the field of mobile communication, 8 kbps at the time of voice encoding for each frame for a speech channel,
20 msec from 4 kbps, 2 kbps and 1 kbps data rate according to the strength of voice energy
A predetermined data rate is selected for each frame of c, and information of the selected data rate is transmitted without notifying the receiving side. On the receiving side, when the CRC added to the 8 kbps and 4 kbps frames and the symbols convolutionally coded for each frame are Viterbi-decoded, the symbols transmitted demodulated based on the obtained information are transmitted most out of the four data rates. Choose a data rate that is likely to have
Then, variable data rate transmission such as outputting decoded data of a corresponding data rate is performed. In such a transmission method, since the data rate information does not have to be sent to the receiving side, the transmission rate can be improved and the control can be simplified.

[0003]

The above-mentioned C
In the DMA cellular telephone system, a data rate per channel is low because voice data is transmitted, and a master station having a speed of about 1000 times the data rate is used.
Using a clock, the receiver decodes all data for the four rates that may have been transmitted for each frame, and then performs processing with a single Viterbi decoder even if rate determination is performed. be able to. However, in the field of mobile communication, services that perform high-speed data transmission are being studied in addition to services that mainly perform voice and low-speed data transmission. Particularly in the case of mobile communication, devices that can increase the data rate are considered. It is desired that the rate of the master clock be as low as possible from the viewpoints of feasibility and power consumption.

When the master clock rate is lowered, the data rate per channel becomes 100 kbp.
In the case of s or more, it becomes difficult to make the ratio between the master clock and the data rate about 1000 times. Then, for all data rates that may have been transmitted, decoding and rate determination of all data on the receiving side with a single Viterbi decoder increases the amount of signal processing, and thus the configuration of the communication device is increased. However, there has been a problem that the speed is increased and the complexity is increased, and its feasibility is difficult.

In view of the above situation, the present invention selects a data rate that is most likely to be transmitted by a Viterbi decoder on the receiving side by selecting a signal transmitted without notifying the data rate information to the receiving side. It is an object of the present invention to provide a variable data rate communication device that can transmit data at a high data rate in variable data rate transmission for decoding.

[0006]

In order to achieve the above object, a variable data rate communication apparatus according to the present invention provides a variable data rate communication apparatus in which one of a plurality of predetermined transmittable data rates is one for each frame. When receiving and decoding a transmission signal that has been selected and is convolutionally coded and transmitted at the selected data rate, a rate determination is performed for all data rates that may have been transmitted. Setting means for setting the number of bits and setting each rate judgment threshold, and performing the Viterbi decoding up to the number of rate judgment bits set by the setting means for each transmittable data rate, thereby obtaining the data rate. Rate determination information generating means for generating rate determination information for each of the plurality of mobile stations, and comparing the generated rate determination information with the rate determination threshold. Rate determining means for determining a data rate most likely to have been transmitted, and based on the data rate determined by the rate determining means, performs Viterbi decoding of frame data after the rate determination bit number. I am trying to do it.

Further, in another variable data rate communication device according to the present invention, one data rate is selected for each frame from a plurality of types of predetermined data rates that can be transmitted, and the selected data rate is selected. Set the number of rate determination bits to perform rate determination multiple times per frame for all data rates that may have been transmitted when receiving and decoding transmission signals transmitted with convolutionally encoded data. And setting means for setting each rate judgment threshold value, and performing Viterbi decoding up to the number of rate judgment bits set by the setting means for each transmittable data rate, so that the rate judgment for each data rate is performed. Rate determination information generating means for generating information; and comparing the generated rate determination information with the rate determination threshold. A rate determination unit that determines a data rate that is likely to be transmitted, the rate determination unit compares the rate determination information with the rate determination threshold set by the setting unit, and transmits the data. If there is a plurality of applicable data rates, the subsequent decoding performs decoding on the plurality of applicable data rates, and the data rate determined by the rate determination unit is reduced to one. The rate determination is sequentially performed with the set number of rate determination bits until the rate determination is completed.

Further, in the variable data rate communication device, the rate determination information includes a symbol error number obtained by re-encoding the Viterbi-decoded decoded data and comparing the received data with the received data, and a maximum likelihood path. It is a path metric quantity.

Further, for all data rates that may have been transmitted, the demodulated signal power,
A line quality estimating unit for estimating the line quality from the history of the rate determination parameter, and when the line quality is estimated to be good by the line quality estimating unit, the rate determination threshold set by the setting unit is transmitted. The number of data rate candidates determined to have a high possibility is set to be small, and when the line quality is estimated to be poor by the line quality estimating means, the rate determination threshold set by the setting means is transmitted. It may be set so that the number of data rate candidates determined to be likely to have increased is increased. In this case, the setting of the rate determination bit number may be changed instead of the rate determination threshold.

Furthermore, if the rate determination means determines that the data rate which may have been transmitted by comparing the rate determination parameter with the rate determination threshold cannot be determined, the subsequent decoding is not performed. Alternatively, rate determination inability information may be output.

Still another variable data rate communication device according to the present invention sets one data rate for each frame among a plurality of predetermined transmittable data rates, and sets the data rate to the set data rate. Allocating means for allocating frame data based on the data rate; convolutional coding means for performing convolutional coding of frame data based on the data rate; and coded data output from the convolutional coding means based on the data rate. A symbol generating means for generating a modulation symbol by performing symbol erasure or symbol division by using a modulation means for modulating a modulation symbol output from the modulation symbol generating means, and demodulating the modulation symbol. Demodulation means for obtaining demodulated symbols by performing
Viterbi decoding means for performing a Viterbi decoding by a Viterbi algorithm by adaptively setting a data rate and the number of decoded data for all data rates at which demodulation symbols output from the demodulating means can be transmitted; Storage means for storing a decoding state and decoded data in the Viterbi decoding means, such as a path metric amount, a branch metric amount for each state corresponding to a path memory length, state transition information, and the like;
The Viterbi-decoded decoded data is re-encoded, a rate determination parameter such as the number of symbol errors obtained by comparing with the demodulated symbol and the path metric amount of the maximum likelihood path is output, and transmitted based on the rate determination parameter. Rate determining means for determining a likely data rate; and output means for outputting decoded data at the determined data rate, wherein the Viterbi decoding means includes a frame for all possible data rates transmitted. For each data rate, decoding is performed up to a preset number of rate determination bits, and the storage unit stores the decoding state and the decoded data for each data rate. The rate determination parameter for all data rates by comparing the determination parameter with the rate determination threshold From among them, the data rate most likely to have been transmitted is determined, the decoding state stored in the storage means is read out for the determined data rate, and the data after the rate determination bit number is decoded, and the data rate for one frame is determined. Receiving means for combining the decoded data before rate determination and the decoded data after rate determination stored in the storage means of the corresponding data rate and outputting as one frame of decoded data after the completion of decoding of the frame data of , Is provided.

Still another variable data rate communication device according to the present invention sets one data rate for each frame among a plurality of predetermined transmittable data rates, and sets the data rate to the set data rate. Allocating means for allocating frame data based on the data rate; convolutional coding means for performing convolutional coding of frame data based on the data rate; and coded data output from the convolutional coding means based on the data rate. A symbol generating means for generating a modulation symbol by performing symbol erasure or symbol division by using a modulation means for modulating a modulation symbol output from the modulation symbol generating means, and demodulating the modulation symbol. Demodulation means for obtaining demodulated symbols by performing
Viterbi decoding means for performing a Viterbi decoding by a Viterbi algorithm by adaptively setting a data rate and the number of decoded data for all data rates at which demodulation symbols output from the demodulating means can be transmitted; Storage means for storing a decoding state and decoded data in the Viterbi decoding means, such as a path metric amount, a branch metric amount for each state corresponding to a path memory length, state transition information, and the like;
The Viterbi-decoded decoded data is re-encoded, a rate determination parameter such as the number of symbol errors obtained by comparing with the demodulated symbol and the path metric amount of the maximum likelihood path is output, and transmitted based on the rate determination parameter. Rate determining means for determining a likely data rate; and output means for outputting decoded data at the determined data rate, wherein the Viterbi decoding means includes a frame for all possible data rates transmitted. For each data rate, decoding is performed up to a preset number of rate determination bits, and the storage unit stores the decoding state and the decoded data for each data rate. The rate determination parameter for all data rates by comparing the determination parameter with the rate determination threshold Determine the data rate that may have been transmitted from the inside, and if there are a plurality of applicable data rates, the subsequent decoding performs decoding on the corresponding plurality of data rates, and the data rate determined by the rate determination unit Until the number of bits becomes one, the rate determination is sequentially performed with the set number of rate determination bits, and the decoding state stored in the storage means is read out for the determined data rate, and data after the rate determination bit number is read. After decoding of one frame of frame data is completed, the decoded data before rate determination and the decoded data after rate determination stored in the storage unit of the corresponding data rate are decoded for one frame. Receiving means for outputting as data.

According to such a variable data rate communication device of the present invention, in a communication channel with good channel quality, the number of rate determination bits is reduced for each frame and all data rates that can be transmitted by one rate determination are used. By judging the data rate that is most likely to have been transmitted from inside, and by decoding at only one data rate after the judgment, the amount of signal processing can be greatly reduced. Therefore, communication at a high data rate can be performed. Also, the variable data rate communication device of the present invention,
In a communication channel where the line quality fluctuates, such as a fading channel, the number of rate determination bits or the rate determination threshold can be adaptively changed according to the line quality, so that the signal processing amount can be reduced while maintaining the communication quality. become able to. Further, the variable data rate communication device of the present invention can reduce the amount of signal processing by outputting rate determination inability information without performing subsequent decoding when rate determination is impossible because rate determination inability can be determined. Will be able to

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In one embodiment of the variable data rate communication device according to the present invention, the variable data rate communication device comprises a transmitter and a receiver. And the transmitter comprises a convolutional encoder and a punctured encoder,
The receiver has a Viterbi decoder. Therefore, first, the principle configurations of the convolutional encoder, the Viterbi demodulator, and the punctured encoder will be described. FIG.
(A) is a circuit block diagram illustrating an example of a configuration of a convolutional encoder. The convolutional encoder has a coding rate r = 1 /
2, the constraint length K = 3. FIG. 2B is a chart showing the relationship among the binary sequence input bits Dn, the output code symbol YnZn, and the convolutional encoder state SR1SR2 in the convolutional encoder shown in FIG. .

The convolutional encoder shown in FIG. 1A has three stages of shift registers SR1, SR2, SR3 and modulo 2 (mod2).
ADD1, ADD2, and ADD3. The input bit Dn is input to the first-stage shift register SR1, and is shifted toward the last-stage shift register SR3 every reference clock. The adder ADD1 adds mod2 of the output of the shift register SR1 and the output of the shift register SR2, and the adder A
DD2 is the sum of the output of the adder ADD1 and the output of the shift register SR3.
The first code symbol Yn is generated by adding d2, and the adder ADD3 includes the shift register SR1 and the shift register Sn.
By adding mod2 to the output of R3, the second code symbol Zn
Has been generated. Note that the code symbol for the 1-bit input bit Dn output from the convolutional encoder is the first symbol.
Is a combination of the code symbol Yn and the second code symbol Zn.

In this convolutional encoder, it is assumed that the contents of the three-stage shift registers SR1, SR2, SR3 are all "0" as an initial state. At this time, when "0" is input as the input bit Dn as shown in FIG. 1B, "00" is generated and output as the code symbol YnZn as shown in FIG. 1B. Next, when “1” is input as the input bit Dn, “1” is input as the code symbol YnZn.
1 "is generated and output. A case where two code symbols are output in response to one-bit input is referred to as a coding rate r =＝. A series of input bits Dn input thereafter is “0011” as shown in FIG.
010100 ”, the code symbol YnZn becomes“ 10
11 11 01 01 00 10 00 10 11 ".

At this time, the first-stage shift registers SR1 and SR2
When the state SR1SR2 with the second-stage shift register SR2 is "00", the state remains "00" (Dn = 0) when a new input bit Dn is input, or the state " 10 ".
(Dn = 1) and the state “01” or the state “ 1 ”
1 ”does not occur. Also, if the state SR1SR2 is “0”
In the case of “1”, when a new input bit Dn is input, the state is “00” (Dn = 0) or “10” (Dn =
The state transits to the state of 1) and does not transit to the state “01” or the state “11”. In addition, state SR1S
When R2 is "10", when a new input bit Dn is input, the state "01" (Dn = 0) or the state "1"
1 ”(Dn = 1), and does not remain in the state“ 10 ”or the state“ 00 ”. Furthermore,
When the state SR1SR2 is “11”, the state transits to the state “01” (Dn = 0) when a new input bit Dn is input, or remains in the state “11” (Dn = 1),
There is no transition to state “00” or state “10”.

As described above, there are two states SR1SR2 to which a transition can be made when the next input bit Dn is input from each state SR1SR2, and a diagram showing this transition mode is the same as the transition diagram shown in FIG. Become. In this figure, a broken arrow indicates a state where the input bit Dn is "0", and a solid arrow indicates a state where the input bit Dn is "1". The 2-bit signal added to the dashed arrow line and the solid arrow line indicates a code symbol that is output when the state changes. Figure 1 above
As another method showing the operation of the convolutional encoder shown in FIG. 2A, a trellis diagram shown in FIG. 2 is known. This trellis diagram corresponds to the chart of FIG.
It shows the same contents as the transition diagram of (c).

In the trellis diagram shown in FIG. 2, the ordinate indicates the state S0 to the state S3, and the abscissa indicates the times t1 and t3.
2, t3... The state in FIG.
If the state of SR1SR2 is “00”, the state is S0, and the state of SR1SR2 is “1”.
"0" is shown as state S1, state SR1SR2 is "01" as state S2, and state SR1SR2 is "11" as state S3. A path indicated by a broken line is a path when the input bit Dn is “0”, and a path indicated by a solid line is a path when the input bit Dn is “1”.

The state transition will be described with reference to this trellis diagram. It is assumed that in the initial state, all the shift registers of the convolutional encoder are reset to the state S0. When input bit Dn is input at time t1, state S0 (code symbol “00”) or state S1
(Code symbol “11”). Next, when the input bit Dn is input at time t2, the state S
In the case of 0, the state transits to either the state S0 (code symbol “00”) or the state S1 (code symbol “11”),
In the case of the state “10”, the state S2 (the code symbol “1”
0 ”) or state S3 (code symbol“ 01 ”). That is, at time t2, any of the four states can be taken.

When the input bit Dn is input at time t3, in the case of the state S0, the state transits to either the state S0 (code symbol "00") or the state S1 (code symbol "11"). , State S1, state S2 (code symbol “10”) or state S3 (code symbol “0”).
1)). In the case of the state S2, the state transits to either the state S0 (code symbol 11) or the state S1 (code symbol "00"). In the case of the state S3, the state S2 (code symbol "01") State S3
(Code symbol “10”). Subsequent time t
At 4, t5,..., The same state transition as at time t3 is performed. In this way, convolutional coding is performed at a coding rate r = 1/2, in which two code symbols are obtained, corresponding to one input bit Dn.

The convolutionally encoded code symbols are modulated and transmitted from the transmitter. The receiver receives the transmitted signal and determines received data. There are two types of this determination, a hard decision and a soft decision. In the case of hard decision, a decision is made based on two quantization levels: “1” if the detection output level is larger than 0, and “0” if the detected output level is smaller than 0. For this reason, the hard decision has a drawback that an error decision occurs at the spread portion of the tail of the probability density function.

On the other hand, the soft decision is a decision method for compensating for the drawbacks of the hard decision, and its configuration is shown in FIG. In the soft decision, received data is input to an A / D converter and N-bit quantized. If the number of bits N is about 3 bits, good characteristics can be obtained at the time of decoding. Therefore, the case of 3-bit quantization, that is, quantization at 8 levels will be described with reference to FIGS. As shown in FIG. 3B, the level difference between the signal “−1” and the signal “1” is equally divided into eight. At this time, one quantization level width is T, and the center level is “0”. FIG. 3C shows a chart of metrics for the code “1” and metrics for the code “0”.

In this table, when a signal a having a level as shown in FIG.
If the code is “1”, the path metric is 3, and
The path metric when the signal a has the code “0” is −3.
Becomes Further, for example, as shown in FIG.
Is received, the path metric when the signal b is set to the code “1” is −1, and the path metric when the signal b is set to the code “0” is 1. In this way, the path metric of each path is determined, and this path metric is used at the time of maximum likelihood decoding for decoding a convolutional code so that optimal decoding is performed.

In this maximum likelihood decoding method, the received sequence is compared with all the possibly transmitted sequences prior to the determination,
A sequence indicating the maximum likelihood value is selected and compared with the received signal. A decoder that realizes such a maximum likelihood decoding method is a Viterbi decoder. The decoding algorithm of the Viterbi decoder will be described with reference to the trellis diagram shown in FIG. 4, but here, for the sake of explanation, it is assumed that the received sequence has undergone a hard decision. The trellis diagram shown in FIG. 4 is a trellis diagram when decoding the code symbol encoded by the convolutional encoder shown in FIG. 1, and is the same as the trellis diagram of the convolutional encoder shown in FIG. Have been.

Here, suppose that the transmitted code symbol is “00, 00, 00, 00, 00”, an error occurs in the transmission path, and the received sequence after hard decision is “01, 00, 00, 00, 00”. , 00, 10, 00, 00 ". At time t1, the reception sequence “01” is input to the Viterbi decoder, and there are two states that can be transitioned from the initial state S0, the state S0 and the state S1, and assuming that the state transits to each state, Check the degree of coincidence of the paths that transit
The degree of coincidence is calculated as “1” when the codeword of the path to the transition state and the received sequence code are the same code, and as “0” when the code is different. For example, assuming that a transition is made to state S0, a match between the codeword "00" of the path to state S0 and the code of the received sequence is calculated. In this case, since only one symbol matches, the matching degree becomes “1”, and the state S0 at time t1
The degree of coincidence is indicated by adding “1” to the top. Further, assuming that a transition is made from the state S0 to the state S1, the code word "11" of the path to the state S1 matches the code "01" of the received sequence, and the degree of matching is "1".

At this time, the calculated degree of coincidence of the state S0 and the degree of coincidence of the state S1 are both "1", and this value is the accumulated metric amount of the states S0 and S1 at time t1. Next, at time t2, the received sequence “00” is input, and the accumulated metric values of the four possible states are obtained. The accumulated metric amount in the state S0 at the time t2 is obtained by adding the coincidence “2” to the accumulated metric amount “3”.
Is obtained. In addition, the accumulated metric amount in the state S1 at the time t2 is maintained at the accumulated metric amount “1” since the degree of coincidence is “0”. Further, the accumulated metric amount in the state S2 at the time t2 is added with the coincidence “1” to obtain the accumulated metric amount “2”. Similarly, the accumulated metric amount in the state S3 is added with the coincidence “1”. Thus, the accumulated metric quantity “2” is obtained.

Subsequently, at time t3, the received sequence "10" is input, and the accumulated metric values of four possible states are obtained. At the time t3, the accumulated metric amount in the state S0 is added with the degree of coincidence "1" to obtain the accumulated metric amount "4". Also, the accumulated metric amount of the state S1 at the time t3 is obtained by adding the coincidence “1” when transitioning from the state S0 to obtain the accumulated metric amount “4”, and is coincident “1” when transitioning from the state S2. Are added to obtain the cumulative metric quantity “3”. At this time,
In the case of a path with a small accumulated metric quantity, it is unlikely that the path is correct. Therefore, as shown in the figure, the path with the metric quantity “3” is truncated, and the path with the high metric quantity “4” is made to survive. You.

Further, the accumulated metric amount of the state S2 at the time t3 is added with the degree of coincidence "2" when transitioning from the state S1 to obtain the accumulated metric amount "3".
When transiting from the state S3, the coincidence degree is set to “0” and the accumulated metric amount becomes “2”. Therefore, as shown in the figure, the path with the accumulated metric amount “2” is discarded, and the path with the accumulated metric amount “3” survives. Similarly, the accumulated metric amount of the state S3 is “0” when the state transitions from the state S1, and the accumulated metric amount is “1” when transitioning from the state S1, and the coincidence degree “2” is added when the state S3 transitions. A metric quantity “3” is obtained. Therefore, the path of the accumulated metric quantity “1” is discarded as shown in FIG.
The path of the accumulated metric amount “3” survives.

At time t4, the received sequence "00" is input, and the same operation as at time t3 is performed. At time t4, the accumulated metric amount in state S0 is "6" and the accumulated metric amount in state S1 is "5". , The accumulated metric amount of the state S2 is “5”, and the accumulated metric amount of the state S3 is “5”. Also, a path with a small accumulated metric amount is discarded. Further, at time t5, the received sequence “00” is input, and the same operation as at time t3 is performed. At time t5, the accumulated metric amount in state S0 is “8”, and the accumulated metric amount in state S1 is “7”. , The accumulated metric amount of the state S2 is “6”, and the accumulated metric amount of the state S3 is “6”. Also in this case, a path with a small accumulated metric amount is discarded.

As described above, the accumulated metric amount indicating the path of the originally transmitted path is the largest accumulated metric amount compared to the accumulated metric amounts in other states. Then, if the path having the largest accumulated metric is followed,
The decoded transmission information sequence D can be obtained. In this case, it suffices to follow the path from the state S0 indicated by the thick broken line in FIG. 4, and the transmission information sequence d = “00, 0”.
0000, 00, 00 "can be obtained. As can be seen from the decoded transmission information sequence D, it can be seen that the error in the reception sequence has been corrected and correct data has been reproduced.

It is to be noted that if a code that periodically sets the state of the convolutional encoder to "00" is input to the convolutional encoder and the code symbol is transmitted, the decoding algorithm in the receiver periodically changes. Since decoding can be performed on the assumption that the state converges to the state S0, paths deriving from the states S1 and S3 do not need to be considered at the time when the state should converge to the state S0. You can narrow down. In addition to the bits after the input bits of a certain block on the transmitting side, the bits whose state converges to S0 are called tail bits. If the constraint length of the convolutional encoder is K, the number of tail bits is (K− 1) Bits whose contents are all "0". In the above-described Viterbi decoder, decoding of a received sequence in the case of performing a hard decision has been described. However, when decoding a received sequence in which a soft decision is made, a transition from one state to two states is made. The metric amount of each path is a quantized metric amount as shown in FIG. 3C, and the accumulated metric amount of the state at each time is a value obtained by accumulating the quantized metric amount of the corresponding path. Become.

By the way, the coding rate r in the above-mentioned convolutional encoder is 1/2, and two code symbols are output for one bit input. Therefore, when the transmission path band is relatively small, This causes inconvenience. In such a case, a punctured code is known as a flexible code that can bring the coding rate close to one. An outline of punctured encoding and its decoding will be described with reference to FIG. In FIG. 5, an input bit Dn at a rate fb is input to a convolutional encoder 201 having a coding rate r = 1/2, and is subjected to convolutional encoding to be a code symbol YnZn at a rate 2fb. This code symbol YnZn is output to the symbol erasing circuit 2
02, m symbols in a code block composed of n symbol code symbols YnZn are periodically deleted. The modulation symbol Wn output from the symbol erasure circuit 202 is a punctured code, and its speed is {2fb
(nm) / n}.

A punctured encoded modulation symbol Wn
Is transmitted from the transmitter to the transmission path and received by the receiver as the symbol Wn '. When decoding the punctured code in the receiver, the symbol periodically deleted in the dummy symbol insertion circuit 203 is inserted, and the speed is 2fb.
And the coding rate r = 1
The transmitted data can be reproduced by being decoded by the Viterbi decoder 204 of / 2. The speed of the decoded data Dn is fb. When performing punctured coding, the number n of symbols forming one block and the number m of erasure symbols in one block can be variously selected, and an arbitrary coding rate can be obtained.

FIG. 5B shows an example of an erasure symbol in the case of punctured coding, and FIG. 5C shows a table of an erasure map with respect to a coding rate. In the example shown in FIG. 5B, the coded output coded at the coding rate r = 1/2 is
This shows an erasure symbol when a punctured code with a coding rate r = ２ is used. Here, the number n of symbols in one block is set to four, and the last symbol of one block is periodically deleted as shown by adding a cross. As a result, a punctured code having a coding rate r = ２ is obtained. Note that at a coding rate r = 2/3, three code symbols are output for two input bits Dn.

FIG. 3C shows a typical erasure map corresponding to the coding rate of the erasure symbol to be periodically erased. The original code is a convolution with the coding rate r = 1/2. Sign. In this erasure map, a symbol that is periodically erased in the block is indicated as “0”. That is, the symbol at the position of “0” in the block is erased. Then, the receiver inserts a dummy symbol at the erased position to obtain a symbol sequence of the original coding rate r = 1/2.

Next, FIG. 6 shows a schematic configuration of a variable data rate transmitting apparatus according to the present invention for transmitting a variable data rate signal in the variable data rate communication apparatus according to the present invention. In FIG. 6, reference numeral 1 denotes a transmission timing setting unit for setting the transmission timing of each block of transmission data in accordance with a data rate input for each frame, 2 an input buffer for storing transmission data, and 3 an addition buffer at the end of the frame. 4 is a frame data generating unit for generating a data sequence to be transmitted for each frame, and 5 is a frame data generating unit 4.
A convolutional coder for performing convolutional coding of data in units of frames output from the CDMA, and performs symbol erasure or symbol division of the code symbols output from the convolutional encoder 5 for each data rate to be transmitted A symbol generation unit 7 that performs punctured coding to generate modulation symbols, and a modulation unit 7 that performs modulation such as DPSK or QPSK on the modulation symbols output from the symbol generation unit 6.

By the way, the frame specification in the variable data rate communication device according to the embodiment of the present invention is, for example, as shown in FIG. As shown in the frame specifications shown in FIG. 11, there are three types of data rates that can be used: rate A, rate B, and rate C. The coding rate (r) is 1/2, and the constraint length (K) is Seven convolutional codes are used. The operation of the transmitting apparatus shown in FIG. 6 at the rate A of coding rate 1/2 will be described. At the rate A, the number of information bits per frame is 506 bits, and all "0" s generated from the tail bit generating unit 3 are set to "0". "
Are set to 6 bits, and a frame composed of a total of 512 bits is generated by the frame data generation unit 4. The information bits of one frame of 512 bits are subjected to convolutional coding (coding rate r = １ ／) in the convolutional encoder 5 to generate 1024 code symbols per frame. At the rate A, the number of modulation symbols is equal to the number of code symbols, and the number of symbols is both 1024 symbols. That is, the punctured coding is not performed in the symbol generation unit 6 at the rate A. Then, the modulation unit 7 modulates, for example, QPSK or the like on the modulation symbols of 1024 symbols per frame, and transmits the modulated symbols. The state of the convolutional encoder 5 having the coding rate r = 1/2 and the constraint length K = 7 can be all set to “0” by the 6 tail bits.

The same operation as that of the rate A is performed at the rate B of the coding rate of 3/4. However, at the rate B, the number of information bits per frame is 762 bits, and the tail of all "0" is set. The bits are 6 bits,
One frame is composed of a total of 768 bits, and is output from the frame data generator 4. The information bits of 768 bits per frame are convolutionally coded by the convolutional encoder 5 at a coding rate of １ ／, thereby generating 1536 code symbols. At the rate B, the symbol generation unit 6 performs punctured coding so that the code rate of 1536 symbols in one frame is 3/4. At this time, the number of code symbols is 1536 and the number of modulation symbols is 1024 in one frame.

Further, at a rate C of a coding rate of 7/8, 1
The number of information bits per frame is 890 bits,
The tail bits of all “0” are 6 bits, and one frame is composed of a total of 896 bits, and is output from the frame data generation unit 4. The convolutional encoder 5 converts the 896 information bits of one frame into 1 /
By performing convolutional encoding in 2, code symbols of 1792 symbols are generated. At the rate C, a coding rate of 7 /
8 is punctured by the symbol generation unit 6 so as to be 8. At this time, the number of code symbols is 1792 and the number of modulation symbols is 1024.

Next, the operation of the transmitting apparatus shown in FIG. 6 for performing variable data rate transmission will be described. When the transmission data rate of the rate A having a coding rate of 1/2 is input to the transmission timing setting unit 1 of the variable data rate transmitting apparatus shown in FIG.
, A timing signal is sent from the transmission timing setting unit 1 to each block of the modulation unit 7. When 506 information bits are input to the input buffer 2, the information bits are sent to the frame data generation unit 4,
After the information bits for the bits, tail bits consisting of all 6 “0” s output from the tail bit generation unit 3 are added, and data for one frame is generated.

The generated data for one frame is supplied to a convolutional encoder 5 and convolutionally encoded. The convolutional encoder 5 performs convolutional coding using a convolutional code having a coding rate r = 1/2 and a constraint length K = 7. When the input and output of the convolutional code are represented by a trellis diagram, the constraint length K is 7, and the number of states is 2 ^K-1 states, that is, 64 states. In the case of the rate A, since the punctured coding and the symbol division are not performed, the symbol generation unit 6
In this case, no processing is performed, and 1024 modulation symbols equal to the number of code symbols are output. The modulation symbol is modulated by the modulation unit 7 and a frame of rate A is output as a transmission signal.

Subsequent to the transmission data rate of rate A of coding rate 1/2, the transmission timing setting section 1 sends the coding rate 3 /
When a transmission data rate of rate B of 4 is input, a timing signal is transmitted from the input buffer 2 to each block of the modulation unit 7 from the transmission timing setting unit 1 according to the frame specification of rate B. Then, 762 is input to the input buffer 2.
When the information bits for the bits are input, the information bits are sent to the frame data generation unit 4 and are composed of all 6 “0” s output from the tail bit generation unit 3 after the 762 information bits. Tail bits are added to generate data for one frame of rate B. The generated data for one frame is supplied to the convolutional encoder 5 and convolutionally encoded. The convolutional encoder 5 performs convolutional encoding using a convolutional code having an encoding rate of 1/2 and a constraint length of 7. Then, in the case of rate B, the coding rate is 3
Punctured encoding is performed by the symbol generation unit 6 so that the number of code symbols of 1536 symbols is 10
The number of modulation symbols is 24 symbols. The modulation symbols of 1024 symbols in one frame are modulated by the modulator 7 and a frame of rate B is output as a transmission signal.

Further, when the transmission data rate of the rate C of 7/8 is input to the transmission timing setting unit 1 following the transmission data rate of the rate B of the coding rate of 3/4, the same as described above. Is repeated from the input buffer 2, 890 information bits are sent to the frame data generation unit 4, and after the 890 information bits, the 6-bit all “ A tail bit consisting of a "0"
Frame data is generated. Then, the convolutional coded data is convolutionally coded by the convolutional coder 5 and the coding rate is 7/8.
Punctured encoding is performed by the symbol generation unit 6 so that the number of encoded symbols of 1792 symbols is 102
The number of modulation symbols is 4 symbols. The modulation symbols of 1024 symbols in one frame are modulated by the modulation unit 7, and a frame of rate C is output as a transmission signal.

The above-described processing is performed according to the transmission data rate input to the transmission timing setting unit 1. For example, frame 1 is rate A, frame 2 is rate B, frame 3 is rate C, and frame 4 is rate B ..., the transmission data is input to the input buffer 2 after inputting 506 bits of the rate A, 762 bits of the rate B,
890-bit data of rate C and 762-bit data of rate B and transmission data of each data rate are sequentially input, and the above-described processing is performed for each rate. The transmission signal at this time is as shown in FIG.

Next, FIG. 7 shows a schematic configuration of a variable data rate receiving apparatus according to the present invention for receiving a variable data rate signal in the variable data rate communication apparatus according to the present invention. In FIG. 7, reference numeral 101 denotes a demodulation unit that performs soft-decision demodulation of a received signal; 102, a symbol buffer that stores the soft-decision demodulation symbols output from the demodulation unit 101 in frame units; Viterbi decoder 104 performs maximum likelihood decoding of demodulated symbols at all possible data rates by the Viterbi algorithm, and 104 decodes the decoded data rate output from Viterbi decoder 103 and the number of decoded bits. The rate judgment is performed from the internal unit information and the rate judgment parameters such as the average value of the number of symbol errors obtained by re-encoding the decoded data at the time of decoding for each frame and comparing with the demodulated symbol and the average value of the path metric amount of the maximum likelihood path. If there is a determined data rate, the decoding data rate and decoding bit count Outputs decoder control information etc., the data rate decision unit that performs output when the rate indeterminable information rate indeterminable.

Reference numeral 105 denotes a data rate determination unit 104
The path metric amount of the maximum likelihood path from the Viterbi decoder 103, the branch metric amount for each state (state) for the path memory length, and the state transition information according to the timing of the decoding data rate in the decoder control information output from the The decoding state storage unit 106 that writes or reads information according to the decoder control information is determined to be the writing of the decoded data output from the Viterbi decoder 103 according to the decoder control information output from the data rate determination unit 104. The decoded data storage unit outputs decoded data decoded at the data rate.

Next, a schematic configuration of the data rate determination section 104 is shown in FIG. 8, reference numeral 111 denotes a symbol error number average value and a maximum likelihood value obtained by re-encoding decoded data for each frame output from the Viterbi decoder 103 and comparing the re-encoded data with the demodulated symbols demodulated by the demodulation unit 101. A rate determination parameter storage unit for storing a rate determination parameter such as an average value of a path metric amount of a path, 11
Reference numeral 2 denotes a threshold setting unit that sets thresholds of rate determination parameters for all data rates that may have been transmitted. Reference numeral 113 denotes a rate determination parameter output from the rate determination parameter storage unit 111 and output from the threshold setting unit 112. A comparison unit 114 that performs comparison with a determination threshold value determines a data rate that may have been transmitted based on the comparison result output from the comparison unit 113, and sends a determination data rate to the decoder control unit 115. If the determination is impossible, a rate determination unit that outputs rate determination inability information; 115, a decoded data rate output from the Viterbi decoder 103;
And a decoder for setting decoder control information such as a decoding data rate and a decoding bit count number to be decoded by the decoder according to the decoder internal information such as the decoding bit count number and the judgment data rate output from the rate judgment unit 114. It is a control unit.

Next, a description will be given of an example in which the rate determination operation of the variable data rate receiving apparatus configured as described above is performed once in one frame in accordance with the rate determination specification 1 shown in FIG. As shown in FIG. 9A, an operation when a transmission signal in which the data rates of frame 1 to frame 4 are set to rate A, rate B, rate C, and rate B, respectively, will be described. The received modulation symbols of frame 1 are input to demodulation section 101 shown in FIG. 7 and subjected to soft-decision demodulation, and the modulation symbols subjected to soft-decision demodulation are stored in symbol buffer 102. After 1024 modulation symbols for one frame are input to the symbol buffer 102, the modulation symbols subjected to soft-decision demodulation are sequentially output to the Viterbi decoder 103. The Viterbi decoder 103 performs decoding by the Viterbi algorithm for the number of rate determination bits shown in FIG. That is, 256 bits, 384 bits, and 448 bits are respectively performed in the order of rate A, rate B, and rate C, and after Viterbi decoding for each bit number is completed,
The decoding state such as the path metric amount of the maximum likelihood path, the branch metric amount for each state corresponding to the path memory length, and the state transition information is written in the decoding state storage unit 105. further,
The decoding data is re-encoded at the time of decoding for each frame, and the rate determination parameters such as the symbol error number average value and the path metric amount average value of the maximum likelihood path obtained by comparing with the demodulated symbols output from the demodulation unit 101 and decoding are determined. Decoder internal information such as the performed data rate and the number of bits to be decoded is output to data rate determining section 104.

The data rate determination section 104 stores the rate determination parameter in the rate determination parameter storage section 111 and sends the decoder internal information to the decoder control section 115. Then, after the decoder internal information for four data rates is input to the decoder control unit 115, the rate determination parameters for three data rates are sent from the rate determination parameter storage unit 111 to the comparison unit 113, and the threshold setting unit 112 Are sent to the comparison unit 113 for three data rates. The comparison result of the two inputted to the comparing section 113 is sent to the rate determining section 114. In the rate determination specification 1 shown in FIG. 13, the number of rate determinations in one frame is one, so the rate determination unit 114 determines the data rate most likely to be transmitted from the comparison result, and The decoder control unit 115 generates a decoding data rate and decoding timing based on the determined data rate, and outputs the generated decoding data rate as decoder control information.

Here, it is assumed that the rate determination is performed as shown in FIG. That is, in frame 1, rate A is determined as the data rate most likely to have been transmitted, so that the data rate determination unit 104 determines from the data rate determination unit 104 that the decoder control information for rate A is the Viterbi decoder 103, The data is output to the state storage unit 105 and the decoded data storage unit 106. Then, the decryption state storage unit 10
Numeral 5 reads out the path metric amount of the maximum likelihood path of the rate A, the branch metric amount for each state corresponding to the path memory length, and the state transition information to the Viterbi decoder 103, and decodes the internal state of the Viterbi decoder 103 into 256 bits of the rate A. The state is reset to the state immediately after, and decoding from the remaining 257th bit to 512 bits of one frame is performed.

The decoded data is sent to the decoded data storage unit 106. The decoded data storage unit 106 decodes the decoded data of the 257th bit to the 512th bit after the decoded data of the rate A from the first bit to the 256th bit. And outputs the data as data for one frame.
Since the above-described decoding operation is also performed sequentially on frames 2 to 4, it is sequentially determined that the frame B has the rate B, the frame 3 has the rate C, and the frame 4 has the rate B, and the decoded data of each rate is stored in the decoded data storage unit. It is output from 106.

Next, an example of an operation in which the rate determination is performed up to three times for each frame in accordance with the rate determination specification 2 shown in FIG. 14 will be described. In this case, as shown in FIG. 10A, it is assumed that a transmission signal in which the data rate of frame 1 to frame 4 is rate A to rate C is received, and the decoding operation is shown in FIG. This will be described with reference to the decoding timing chart shown below. The received modulation symbols of frame 1 are input to demodulation section 101 shown in FIG. 7 and subjected to soft-decision demodulation, and the modulation symbols subjected to soft-decision demodulation are stored in symbol buffer 102. In the symbol buffer 102, the number of modulation symbols 102 for one frame
After the four symbols are input, the modulated symbols subjected to soft-decision demodulation are sequentially output to the Viterbi decoder 103. The Viterbi decoder 103 performs decoding by the Viterbi algorithm in FIG.
Is performed for the number of rate determination bits for each data rate shown in FIG.

That is, 192 bits, 288 bits, and 336 bits are respectively performed in the order of rate A, rate B, and rate C in accordance with the specification of rate determination 1, and after completion of Viterbi decoding for each bit number, the path metric of the maximum likelihood path is determined. The decoding state such as the amount, the branch metric amount for each state corresponding to the path memory length, and the state transition information is written in the decoding state storage unit 105. Furthermore, at the time of decoding for each frame, the decoded data is re-encoded, and a rate determination parameter such as a symbol error number average value and a path metric amount average value of the maximum likelihood path obtained by comparing with a demodulated symbol output from the demodulation unit 101 and Decoder internal information such as the decoded data rate and the number of decoded bits is output to data rate determining section 104. As a result, in rate determination 1 of frame 1, it is determined that the possibility of transmission at the data rate of rate A is low, and the decoder control information of rate B and rate C is transmitted from data rate determination section 104 to the Viterbi decoder. 10
3. Output to the decoding state storage unit 105 and the decoded data storage unit 106.

Thus, the rate judgment 2 based on the rate B and the rate C is continuously performed. That is, the decoding state storage unit 105 reads out the path metric amount of the maximum likelihood path at the rate B, the branch metric amount for each state corresponding to the path memory length, and the state transition information to the Viterbi decoder 103 to store the internal state of the Viterbi decoder 103. The state immediately after the decoding of the 288 bits of the rate B is reset, and the decoding of the rate B from the 289th bit to the 576 bits is performed. Then, the path metric amount of the maximum likelihood path at the rate C, the branch metric amount for each state corresponding to the path memory length, and the state transition information are read out to the Viterbi decoder 103 and read out by the Viterbi decoder 1.
03 is reset to the state immediately after the decoding of the 336 bits of the rate C, and the rate C is decoded from the 337th bit to the 672th bit.

After completion of decoding for each bit number, the path metric amount of the maximum likelihood path, the branch metric amount for each state corresponding to the path memory length, and the state transition information are written in the decoding state storage unit 105, and the The decoding data is re-encoded and the decoded data is compared with the demodulated symbols. The symbol error number average value and the rate metric parameter such as the average value of the path metric amount of the maximum likelihood path are obtained, and the decoded data rate and the number of decoded bits. Is output to the data rate determination unit 104. In the data rate determination unit 104, based on the given information, a rate determination is made as to which data rate is more likely to have been transmitted for the rate B and the rate C, and as shown in FIG. In frame 1, in rate determination 2 for the second time, rate C is determined as the data rate that is most likely to have been transmitted.

Based on this determination, the decoder control information of the rate C is transmitted from the data rate determination unit 104 to the Viterbi decoder 1.
03, decryption state storage unit 105, decryption data storage unit 106
Is output to The decoding state storage unit 105 reads out the path metric amount of the maximum likelihood path at the rate C, the branch metric amount for each state corresponding to the path memory length, and the state transition information to the Viterbi decoder 103, and reads out the Viterbi decoder 103.
The internal state is reset to the state immediately after the decoding of the rate C of 672 bits, and decoding from the 673th bit to 896 bits is performed. This decoded data is sent to the decoded data storage unit 106, and the decoded data storage unit 106 outputs 67 bits from the first bit.
By adding decoded data from the 673th bit to 896 bits after the decoded data of the rate C up to 2 bits,
Output as data for one frame.

When the decoding operation for frame 1 is completed, the above-described decoding operation is performed for frame 2 as well. In frame 2, rate A is determined as the data rate most likely to have been transmitted by the above-described rate determination 1 and rate determination 2. Is determined, and the decoded data at the rate A is output from the decoded data storage unit 106. When the decoding of the frame 2 is completed, the decoding operation of the frame 3 is performed. In the frame 3, even if the second rate determination 2 is performed, two data rates within the rate determination threshold remain as the rate B and the rate C. Become like In this case, 3 in rate B and rate C
The third rate determination 3 is performed, and the rate determination 3 determines that the rate B is the data rate most likely to have been transmitted. Therefore, the decoded data of one frame at the rate B is output from the decoded data storage unit 106. Furthermore, in frame 4, since there is no rate within the rate determination threshold in the first rate determination, the rate determination is disabled, and the data rate determination unit 104
Output rate inability information. In this case, the subsequent rate determination operation is not performed.

Next, FIG. 15 shows an example of a circuit block diagram showing a configuration for re-encoding the decoded data, which is one of the rate determination parameters, and obtaining the number of symbol errors obtained by comparing the decoded data with the demodulated symbols. In FIG. 15, the demodulated symbols demodulated by the demodulation unit are decoded by the Viterbi decoder 120 using the Viterbi algorithm, and decoded data is output. This decoded data is supplied to the encoder 121 and encoded into a convolutional code having the same coding rate and constraint length as the transmitting device. The coded parallel symbols output from the encoder 121 are converted into serial data by a parallel / serial (P / S) converter 122 and supplied to the mismatch detector 123. Further, the demodulated symbol is supplied to the Viterbi decoder 120 through P
/ S converter 122 is delayed by the operation delay time, and P
The symbol output from the / S converter 122 and the decoded symbol are set at the same time, and the
3 is supplied.

The mismatch detector 123 detects a mismatch between both input symbols and outputs symbol error number information. For example, if the symbol output from P / S converter 122 is “00100011” and the demodulated symbol is “01000011”, the second and third symbols
Since the second symbol does not match, the information of “2” as the number of symbol errors is output from the mismatch detector 123. Further, the detection of the mismatch is not performed for the lost symbol due to the punctured coding.

Simulation results of the average value (path metric) of the maximum likelihood path versus the average value of the number of symbol errors (symble error) at each data rate when a transmission signal is decoded at an arbitrary transmittable data rate. 16 to 18 are shown in FIGS. FIG. 16 (a)
Is that the number of frames is 10678 frames, the bit error rate (BER) is about 10 ^-5 , and the coding rate r = 1 of the transmitting device.
/ 2 (data rate r = １ ／) and the decoding data rate r = １ ／ of the Viterbi decoder, the path metric average value of the maximum likelihood path and the symbol error count average value (symble) error). FIG. 3B shows that the decoding data rate r of the Viterbi decoder is 3/4.
Average path metric value of the maximum likelihood path (path
(c) shows the distribution of the average value of the maximum likelihood path when the decoding data rate r of the Viterbi decoder is 7/8. The distribution of the value (path metric) and the average value of the number of symbol errors (symble error) are shown.

When observing the distribution of the path metric average value (path metric) and the symbol error number average value (symble error) of the maximum likelihood path shown in FIGS. 16 (a) to 16 (c), the data rate of the transmitting apparatus is shown. (1/2) and the decoded data rate (1/2) match (see FIG. 16A)
Is that the variance of the symbol error number average value and the variance of the path metric amount average value of the maximum likelihood path are respectively small, but when the data rate of the transmitting apparatus does not match the decoded data rate, the symbol error number average value and the maximum likelihood path It can be seen that the variance of each of the path metric amount average values increases. Utilizing this, as shown in FIG. 19, when the average value of the number of symbol errors is equal to or less than Tsym1 and the average value of the path metric amount of the maximum likelihood path is equal to or less than Tpm1, the decoded data rate at that time is the highest. The range A is determined to be a data rate that is highly likely to have occurred, and the average number of symbol errors is Tsym2
When the average value of the path metric amount of the maximum likelihood path is Tp
A range B of m2 or more is a range B at which the decoded data rate at that time is determined to be a data rate having a low possibility of being transmitted. Such a determination is made by the data rate determination unit 104.

That is, in the block diagram of the data rate determination section shown in FIG. 8, when the rate A of the coding rate １ ／ is determined, the threshold The value Tsym1, the average value Tpm1 of the path metric amount of the maximum likelihood path, the average value Tsym2 of the number of symbol errors, which is the threshold of the range B, and the average value Tpm2 of the path metric amount of the maximum likelihood path are set. These set thresholds, the average number of symbol errors in the rate determination parameter written in the rate determination parameter storage unit 111,
The comparison unit 113 compares the average value of the path metric quantity of the maximum likelihood path with the average value, and compares the comparison result with the rate determination unit 1.
14 is output. If the received comparison result is within the range A, the rate determination unit 114 determines that the transmitted data rate is the rate A (coding rate Ａ), and if it belongs to the range B, the data rate is Rate A (coding rate 1 /
It is determined that the possibility of transmission in 2) is low.

FIG. 17A shows that the number of frames is 10678.
Frame, bit error rate (BER) is about 10 ^-5 ,
The coding rate r = 3/4 (data rate r = 3 /
4), the average value of the path metric amount of the maximum likelihood path (path m) when the decoding data rate r = タ of the Viterbi decoder is set to
etric) and the symbol error count average value (symble error). FIG. 3B shows the distribution of the average value of the path metric amount (path metric) of the maximum likelihood path and the average value of the number of symbol errors (symble error) when the decoding data rate r of the Viterbi decoder is 3/4. FIG. 9C shows the average value of the path metric amount (path m) of the maximum likelihood path when the decoding data rate r of the Viterbi decoder is set to 7/8.
etric) and the symbol error count average value (symble error).

Observing the distribution of the average value of the path metric amount (path metric) and the average value of the number of symbol errors (symble error) of the maximum likelihood path shown in FIGS. 3/4) and the decoded data rate (coding rate 3/4) match (FIG. 17).
(B) reference coding rate), the variance of the symbol error number average value and the variance of the path metric amount average value of the maximum likelihood path are respectively small, but when the data rate of the transmitting apparatus does not match the decoded data rate, It can be seen that the respective variances of the symbol error number average value and the path metric amount average value of the maximum likelihood path increase. Therefore, a threshold for setting only the range of the distribution shown in FIG.
By setting (a) and a threshold for setting the range of the distribution shown in FIG.
If the comparison result is within range A, the data rate is rate B
It is determined that the data has been transmitted at (coding rate 3/4), and if it belongs to range B, the data rate is rate B (coding rate 3/4).
It can be determined that the possibility of transmission is low.

FIG. 18A shows that the number of frames is 10678.
Frame, bit error rate (BER) is about 10 ^-5 ,
Coding rate r = 7/8 (data rate r = 7 /
8), the average value of the path metric amount of the maximum likelihood path (path m) when the decoding data rate r = タ of the Viterbi decoder is set to
etric) and the symbol error count average value (symble error). FIG. 3B shows the distribution of the average value of the path metric amount (path metric) of the maximum likelihood path and the average value of the number of symbol errors (symble error) when the decoding data rate r of the Viterbi decoder is 3/4. FIG. 9C shows the average value of the path metric amount (path m) of the maximum likelihood path when the decoding data rate r of the Viterbi decoder is set to 7/8.
etric) and the symbol error count average value (symble error).

Observing the distribution of the average value of the path metric amount (path metric) and the average value of the number of symbol errors (symble error) of the maximum likelihood path shown in FIGS. When the coding rate 7/8) and the decoded data rate (coding rate 7/8) match (FIG. 1).
8 (c)), the variances of the symbol error number average value and the path metric amount average value of the maximum likelihood path become smaller, respectively. However, if the data rate of the transmitting apparatus does not match the decoded data rate, the symbol error number It can be seen that the respective variances of the average value and the average value of the path metric amount of the maximum likelihood path increase. Therefore, the threshold setting unit 112 sets a threshold for setting only the range of the distribution shown in FIG. 18C as the range A and a threshold for setting the range of the distribution shown in FIGS. 18A and 18B as the range B. When the comparison result is within the range A, it is determined that the data rate is transmitted at the rate C (coding rate 7/8). When the comparison result is within the range B, the data rate is determined at the rate C (coding rate). 7/8), it is possible to determine that the possibility of transmission is low.

By the way, one of the three data rates is
Path metric amount average value of the maximum likelihood path when one data rate is selected, convolutionally coded, and Viterbi decoded, and
The number of symbol errors obtained by re-encoding the decoded data at the time of decoding and comparing the decoded data with the demodulated symbols takes a smaller value as the channel quality improves, even when decoding at a correct rate. However, when decoding at an incorrect rate, each value takes a substantially constant large value regardless of the line quality.

FIG. 20 shows that the number of data bits per frame when the coding rate is 1/2 is 128 bits and the coding rate r =
、, r =＝, r = ７, and when the number of modulation symbols per frame is set equal,
Frame error rate characteristics 1 for each of the three data rates when rate determination is performed according to the rate determination principle described above.
And a computer simulation result of the rate determination error rate characteristic 1 are shown. Also, if the number of data bits per frame at the coding rate of 1/2 is 256 bits in FIG. 21, the frame of each of the three types of data rates when the rate determination is performed similarly in accordance with the rate determination principle described above. The computer simulation results of the error rate characteristic 2 and the rate determination error rate characteristic 2 are shown.

20 and 21 show the energy / noise power density per bit (Eb / No) versus error rate characteristics under white Gaussian noise, and the frame error rate is shown by a solid line. And the rate determination error rate is shown by the dashed line. Generally, when the rate determination error rate is lower than the frame error rate, the error rate is low, so that the possibility that the rate determination error causes deterioration of communication quality is reduced. Therefore, the rate determination error rate is set lower than the frame error rate.
The communication quality is preferable to perform decoding in a range that is governed by the frame error rate that. Further, the characteristic shown in FIG. 21 is lower than the frame error rate in the degraded Eb / No than the characteristic shown in FIG. It can be seen that the characteristics of the rate determination error rate are improved. In addition, the rate judgment is performed with as few bits as possible,
Reducing the number of data rate candidates can reduce the amount of signal processing required for decoding per frame.

From these facts, it is necessary to reduce the number of rate determination bits and to narrow down data rate candidates by a small number of determinations in a communication channel with high Eb / No and high line quality. This is an effective method to reduce the amount. On the other hand, in a communication channel where the line quality fluctuates such as a fading channel, it is possible to adaptively change the number of rate determination bits or the rate determination threshold according to the line quality.
This is an effective method for reducing the amount of signal processing required for decoding per frame. When estimating the line quality
From the history of the demodulated signal power and rate judgment parameters.
The line quality may be estimated.

[0072]

As described above, the variable data rate communication device according to the present invention comprises:
With the above configuration, in a communication channel with good line quality, the number of rate determination bits may be reduced for each frame, and the possibility of transmission from all data rates that can be transmitted in one rate determination may be determined. By determining the highest data rate and decoding only one data rate after the determination, the amount of signal processing can be greatly reduced. Therefore, communication can be performed at a high data rate. In a communication channel such as a fading channel where the line quality fluctuates, the signal processing amount can be reduced while maintaining the communication quality by adaptively changing the number of rate determination bits or the rate determination threshold according to the line quality. Become like

Further, the variable data rate communication device of the present invention can determine that the rate cannot be determined. If the rate cannot be determined, the probability that a decoding error has occurred even if the rate is correctly determined is determined. And communication quality cannot be guaranteed. Therefore, the entire signal processing amount is reduced by outputting rate determination impossible information without performing subsequent decoding.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an example of a configuration of a convolutional encoder, a table illustrating a relationship between input bits, code symbols, and states of the convolutional encoder, and a state transition diagram of the convolutional encoder.

FIG. 2 is a diagram illustrating an example of a trellis diagram of a convolutional encoder.

FIG. 3 is a block diagram illustrating an example of a configuration for performing soft decision, and a diagram for describing soft decision.

FIG. 4 is a trellis diagram showing a decoding algorithm in a Viterbi decoder.

FIG. 5 is a diagram illustrating an example of a configuration for performing punctured coding, a configuration for decoding a punctured code, and a punctured code.

FIG. 6 is a block diagram illustrating a schematic configuration of a variable data rate transmission device according to the variable data rate communication device of the present invention.

FIG. 7 is a block diagram showing a schematic configuration of a variable data rate receiving device according to the variable data rate communication device of the present invention.

FIG. 8 is a diagram showing a schematic configuration of a data rate determination unit in the variable data rate receiving device of the present invention.

FIG. 9 is a diagram showing a variable data rate transmission example 1 in the variable data rate communication device of the present invention, and a diagram showing a decoding timing chart example 1.

FIG. 10 is a diagram illustrating a variable data rate transmission example 2 in the variable data rate communication device of the present invention, and a diagram illustrating a decoding timing chart example 2;

FIG. 11 is a table showing frame specifications in the variable data rate communication device of the present invention.

FIG. 12 is a chart showing an example of rate determination in the variable data rate communication device of the present invention.

FIG. 13 is a table showing rate determination specification 1 in the variable data rate communication device of the present invention.

FIG. 14 is a chart showing rate determination specification 2 in the variable data rate communication device of the present invention.

FIG. 15 is a block diagram illustrating a configuration example for obtaining symbol error number information among rate determination parameters in the variable data rate communication device of the present invention.

FIG. 16 is a diagram illustrating the relationship between the average value of the path metric amount of the maximum likelihood path and the average value of the number of symbol errors when a transmission signal having a coding rate of で is decoded at three data rates in the variable data rate communication device according to the present invention. It is a figure showing distribution.

FIG. 17 is a diagram illustrating the relationship between the average value of the path metric amount of the maximum likelihood path and the average value of the number of symbol errors when a transmission signal having a coding rate of 3/4 is decoded at three types of data rates in the variable data rate communication device according to the present invention. It is a figure showing distribution.

FIG. 18 is a diagram illustrating the relationship between the average value of the path metric amount of the maximum likelihood path and the average value of the number of symbol errors when a transmission signal having a coding rate of 7/8 is decoded at three types of data rates in the variable data rate communication device according to the present invention. It is a figure showing distribution.

FIG. 19 is a diagram illustrating a setting example of a rate determination threshold in the variable data rate communication device of the present invention.

FIG. 20: Eb / No vs. data rate 1/2, 3/4,
It is a figure which shows the frame error rate of the data rate of 7/8, and the rate determination error rate characteristic 1.

FIG. 21: Eb / No vs. data rate 1/2, 3/4,
It is a figure which shows the frame error rate of 7/8, and the rate determination error rate characteristic 2.

[Explanation of symbols]

 Reference Signs List 1 transmission timing setting unit 2 input buffer 3 tail bit generation unit 4 frame data generation unit 5 convolutional encoder 6 symbol generation unit 7 modulation unit 101 demodulation unit 102 symbol buffer 103 Viterbi decoder 104 data rate determination unit 105 decoding state storage unit 106 Decoded data storage unit 111 Rate determination parameter storage unit 112 Threshold setting unit 113 Comparison unit 114 Data rate determination unit 115 Decoder control unit 116 Transmission timing setting unit 120 Viterbi decoder 121 Encoder 122 P / S converter 123 Non-coincidence detector 124 Delay Means 201 Convolutional encoder 202 Symbol erasure circuit 203 Dummy symbol insertion circuit 204 Viterbi decoder

Claims (8)

(57) [Claims] 1. A transmission signal in which one data rate is selected for each frame from a plurality of predetermined types of data rates that can be transmitted and data of the selected data rate is convolutionally encoded and transmitted. When receiving and decoding, for all data rates that may have been transmitted,
Setting means for setting the number of rate judgment bits for performing the rate judgment and setting each rate judgment threshold; and performing Viterbi decoding up to the number of rate judgment bits set by the setting means for each transmittable data rate A rate determination information generating means for generating rate determination information for each data rate; and comparing the generated rate determination information with the rate determination threshold to obtain a data rate most likely to have been transmitted. Variable data rate communication characterized by comprising: Viterbi decoding of frame data after the number of rate determination bits based on the data rate determined by the rate determination means. apparatus. 2. One of a plurality of predetermined types of data rates that can be transmitted is selected for each frame, and a data signal of the selected data rate is convolutionally coded and transmitted. When receiving and decoding, for all data rates that may have been transmitted,
Setting means for setting the number of rate judgment bits for performing a plurality of rate judgments for each frame and setting each of the rate judgment thresholds; and the number of rate judgment bits set by the setting means for each transmittable data rate. A rate determination information generating means for generating rate determination information for each data rate by performing Viterbi decoding up to the above, and comparing the generated rate determination information with the rate determination threshold to transmit the rate determination information. Rate determination means for determining a data rate with high possibility, the rate determination means comparing the rate determination information with the rate determination threshold set by the setting means, and may have been transmitted. The data rate is determined, and if there are a plurality of applicable data rates, subsequent decoding is performed for the plurality of applicable data rates. There, the rate determination unit data rate to be determined by the variable data rate communication apparatus being characterized in that to perform the sequential rate determined by the number of rate decision bits the set until one. 3. The rate determination information is the number of symbol errors obtained by re-encoding Viterbi-decoded decoded data and comparing it with received data, and the path metric amount of a maximum likelihood path. The variable data rate communication device according to claim 1 or 2, wherein: 4. A line quality estimating means for estimating a line quality from a demodulated signal power and a history of a rate determination parameter for each frame for all data rates that may have been transmitted. If the quality is estimated to be good, the rate determination threshold set by the setting means is set so that the number of data rate candidates determined to have a high possibility of being transmitted is reduced, When the line quality is estimated to be poor, the rate determination threshold set by the setting means is set so that the number of data rate candidates determined to be highly likely to have been transmitted is increased. Claim 1
Alternatively, the variable data rate communication device according to 2. 5. A line quality estimating means for estimating a line quality from a demodulated signal power and a history of a rate determination parameter for each frame with respect to all data rates that may have been transmitted. If the quality is estimated to be good, the number of rate determination bits set by the setting means is set so as to be small, and a rate determination is performed.If the line quality is estimated to be poor by the line quality estimating means, 3. The variable data rate communication device according to claim 1, wherein the rate determination is performed by setting the number of rate determination bits set by the setting unit to be large. 6. When the rate determination means determines that it is impossible to determine a data rate that may have been transmitted by comparing the rate determination parameter with the rate determination threshold, the subsequent decoding is not performed. 3. The variable data rate communication device according to claim 1, wherein the device outputs rate indetermination information. 7. An allocating means for setting one data rate for each frame among a plurality of predetermined transmittable data rates, and allocating frame data based on the set data rate. Convolutional coding means for performing convolutional coding of frame data based on the above, and generating modulated symbols by performing symbol erasure or symbol division on the coded data output from the convolutional coding means based on the data rate. A transmitting unit including a modulation symbol generating unit; a modulating unit that modulates a modulation symbol output from the modulation symbol generating unit; a demodulating unit configured to demodulate the modulation symbol to obtain a demodulated symbol; The demodulated symbols output from the means for all data rates that can be transmitted Viterbi decoding means for adaptively setting the data rate and the number of decoded data to perform Viterbi decoding by the Viterbi algorithm, the path metric amount of the maximum likelihood path, the branch metric amount for each state corresponding to the path memory length, and state transition information Storage means for storing the decoding state and the decoded data in the Viterbi decoding means, etc., and the number of symbol errors and the path metric amount of the maximum likelihood path obtained by re-encoding the decoded data subjected to Viterbi decoding and comparing with the demodulated symbols. And the like, which outputs a rate determination parameter such as, and a rate determination means for determining a data rate having a high possibility of being transmitted based on the rate determination parameter, and an output means for outputting decoded data of the determined data rate. The Viterbi decoding means is provided for every data rate that may have been transmitted for each frame. For each data rate, decoding is performed up to a predetermined number of rate determination bits, the storage means stores the decoding state and the decoded data for each data rate, and the rate determination means The determination parameter is compared with the rate determination threshold to determine a data rate most likely to be transmitted from among the rate determination parameters of all data rates, and the determined data rate is stored in the storage means in the storage unit. The state is read out, the data after the rate determination bit number is decoded, and after the decoding of the frame data for one frame is completed, the decoded data before the rate determination stored in the storage means of the corresponding data rate is determined. Receiving means for combining the subsequent decoded data and outputting as one frame of decoded data; Variable data rate communication and wherein the obtaining. 8. An allocating means for setting one data rate for each frame among a plurality of predetermined transmittable data rates and allocating frame data based on the set data rate. Convolutional encoding means for performing convolutional encoding of frame data based on the above, and generating modulated symbols by performing symbol erasure or symbol division on the encoded data output from the convolutional encoding means based on the data rate. A transmitting unit including a modulation symbol generating unit; a modulating unit that modulates a modulation symbol output from the modulation symbol generating unit; a demodulating unit configured to demodulate the modulation symbol to obtain a demodulated symbol; The demodulated symbols output from the means for all data rates that can be transmitted Viterbi decoding means for adaptively setting the data rate and the number of decoded data to perform Viterbi decoding by the Viterbi algorithm, the path metric amount of the maximum likelihood path, the branch metric amount for each state corresponding to the path memory length, and state transition information Storage means for storing the decoding state and the decoded data in the Viterbi decoding means, etc., and the number of symbol errors and the path metric amount of the maximum likelihood path obtained by re-encoding the decoded data subjected to Viterbi decoding and comparing with the demodulated symbols. And the like, which outputs a rate determination parameter such as the above, determines a data rate having a high possibility of being transmitted based on the rate determination parameter, and output means for outputting decoded data of the determined data rate, The Viterbi decoding means is provided for every data rate that may have been transmitted for each frame. For each data rate, decoding is performed up to a predetermined number of rate determination bits, the storage means stores the decoding state and the decoded data for each data rate, and the rate determination means Determine the data rate that may have been transmitted from the rate determination parameters of all data rates by comparing the determination parameter and the rate determination threshold, if there are multiple corresponding data rates,
Subsequent decoding is performed for a plurality of applicable data rates, and the rate judgment is performed sequentially with the set number of rate judgment bits until the data rate judged by the rate judgment means becomes one. The decoding state stored in the storage means is read out for the data rate obtained, the data after the number of rate determination bits is decoded, and after the decoding of the frame data for one frame is completed,
Receiving means for combining the decoded data before the rate determination and the decoded data after the rate determination stored in the storage means of the corresponding data rate, and outputting as one frame of decoded data. Variable data rate communication device.