JPH0746146A - Decoder - Google Patents

Abstract

PURPOSE:To discriminate the data rate of each unknown frame in a traffic channel with a simple constitution by discriminating the data rate of the traffic channel based on the output of a Viterbi decoding part. CONSTITUTION:A normarization calculation part 21 normalizes the number of symbol errors of each data rate. An average value calculation part 22 calculates an average value of effective path metric volumes in accordance with the path metric volume of one frame of each data rate. Threshold tables corresponding to Eb/No ((energy of bit unit)/noise) for discrimination of four kinds of data rates obtained from the experiment result are stored in a threshold setting part 23. One of stored threshold tables is outputted to a threshold storage part 27 in accordance with Eb/No outputted from an Eb/No estimating part 26 and is stored there. A comparison part 24 compares the normalized number of symbol errors and the path metric volume average value with each threshold, and a discriminating part 25 discriminates the data rate by the comparison results.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a decoding device suitable for use in a receiving portion of a cellular telephone or the like to which a CDMA system, which is one of mobile radio communication systems, is applied.

[0002]

2. Description of the Related Art A method of decoding a convolutionally coded signal using a so-called Viterbi algorithm has been implemented.

That is, in this decoding method, the branch metric for one code block is calculated from the received symbol.
Then, a new survivor path is selected based on this branch metric and the path metric of the survivor path in each state in the pre-decoding step. In addition, the path metric for the new survivor path for each state and the corresponding encoder input data sequence are stored in the path memory.

On the other hand, from the survivor paths of each state, the one with the highest likelihood at that time is selected. Then, the encoder input bits at the time point traced back by the path memory length along the selected path of the stored encoder input data series are output as decoded data. Decoding is performed using such a decoding algorithm.

[0005]

By the way, in a communication system to which a CDMA system or the like used for mobile radio communication is applied, for example, transmission data is convolutionally encoded and transmitted. In recent years, as such a communication system, data is continuously transmitted without adding a tail bit to each frame, and a sync channel (first transmission channel) whose data rate is fixed and a frame unit It has been proposed that data is transmitted in bursts in frame units with a tail bit attached, and that the data channel has a variable traffic channel in frame units (second transmission channel).

In this communication system, the fixed data rate of the sync channel is determined in advance, so that the data rate is known on the receiving side. On the other hand, regarding the traffic channel, a method has been proposed in which data is transmitted without notifying the variable data rate. According to this method, the receiving side can determine the data rate for each unknown frame of the traffic channel. , It becomes necessary to determine the decoding data rate.

However, no suitable method has heretofore been proposed for determining such a decoding data rate.

Even when the data rate of each frame of the traffic channel is transmitted, the data rate is not always correct. That is,
For example, it is conceivable that an error will occur during transmission or decoding, which will cause data to be decoded at an incorrect data rate. Therefore, even in this case, it has been desired that the receiving side discriminate the data rate of each frame of the traffic channel.

The present invention has been made in view of such a situation, and it is possible to determine the data rate at the time of decoding by determining the data rate of each unknown frame of the traffic channel with a simple configuration. To do so.

[0010]

A decoding device according to claim 1 is a decoding device having a Viterbi decoding unit 1 as maximum likelihood decoding means for performing maximum likelihood decoding of received data based on a Viterbi algorithm. The received data is composed of a fixed first channel such as a sync channel and a variable second channel such as a traffic channel having a data rate, and based on the output of the Viterbi decoding unit 1, Data rate determination unit 10 as rate determination means for determining the data rate
(Comparison unit 24 and determination unit 25) or 32 (Comparison determination unit 41) is provided.

The decoding device according to the second aspect further comprises a symbol error number calculation unit 6 as an error number calculation means for calculating the number of symbol errors based on the output of the Viterbi decoding unit 1, and the Viterbi decoding unit 1 One of the outputs of the data rate determination unit 10 (comparison unit 24
And the determination unit 25) or 32 (comparison determination unit 41),
The data rate of the traffic channel is determined based on the number of symbol errors and the path metric amount.

According to a third aspect of the decoding apparatus of the present invention, the traffic channel includes an error detection code, and CRC error detection as error detection means for performing error detection based on the error detection code output from the Viterbi decoding unit 1. The data rate determination unit 10 (comparison unit 24 and determination unit 25) or 32 (comparison determination unit 41) further includes a unit 3 and the CRC error detection unit 3 in addition to the number of symbol errors and the path metric amount.
The data rate of the traffic channel is determined on the basis of the error detection result of 1.

A decoding device according to a fourth aspect of the present invention is a threshold value setting means for setting a threshold value for the number of symbol errors or a threshold value for a path metric amount for each data rate that a traffic channel can take, based on the output of the Viterbi decoding unit 1. Further comprising a threshold value setting unit 23 or 42, the comparing unit 24 and the judging unit 25, or the comparing and judging unit 41 compares the threshold value with the symbol error number and the path metric amount, and based on the comparison result, the traffic It is characterized by determining the data rate of the channel.

The decoding device according to the fifth aspect is characterized in that the threshold setting unit 23 or 42 sets the threshold value based on the output of the Viterbi decoding unit 1 corresponding to the sync channel.

The decoding apparatus according to the sixth aspect is characterized in that the threshold setting unit 23 or 42 sets the threshold value based on the output of the Viterbi decoding unit 1 corresponding to the traffic channel.

In the decoding device according to the seventh aspect, the threshold setting unit 23 sets the threshold for each reliability, and the comparing unit 2
4 and the determination unit 25 sequentially compare the number of symbol errors and the path metric amount with a threshold value having a high reliability and one having a higher data rate from the one having a lower data rate, and the threshold value having the next lower reliability by one. It is characterized in that the data rates are sequentially compared with those with a higher data rate.

In the decoding device according to the eighth aspect, the threshold setting unit 23 sets the threshold for each reliability, and the comparing unit 2
4 and the judgment unit 25 judge the line disconnection standard estimated when the data rate of the traffic channel is judged to be worse than the line disconnection standard estimated when the threshold value setting unit 23 sets the threshold value, the traffic channel data. It is characterized in that the reliability of the threshold applied to determine the rate is lowered and then output.

In the decoding device according to the ninth aspect, the comparison / determination unit 41 compares the number of symbol errors and the path metric amount based on the error detection result of the CRC error detection unit 3, and the traffic channel is used. It is characterized in that the order of thresholds set for each data rate to be obtained is determined.

In the decoding device according to the tenth aspect, the comparison / determination unit 41 compares the number of symbol errors and the path metric amount in order from the threshold value corresponding to the data rate that the traffic channel is likely to take. Characterize.

[0020]

In the decoding device according to the first aspect, the Viterbi decoding unit 1 performs maximum likelihood decoding of the received data based on the Viterbi algorithm. Then, the data rate of the traffic channel is determined based on the output of the Viterbi decoding unit 1. Therefore, the data rate can be easily determined.

In the decoding device according to claim 2,
The number of symbol errors is calculated based on the output of the Viterbi decoding unit 1, and the data rate of the traffic channel is calculated based on the number of symbol errors and the path metric amount which is one of the outputs of the Viterbi decoding unit 1. judge. Therefore, it is possible to accurately determine the data rate.

In the decoding device according to claim 3,
The traffic channel includes an error detection code, and error detection is performed based on the error detection code output from the Viterbi decoding unit 1. Then, the data rate of the traffic channel is determined based on the error detection result, the number of symbol errors, and the path metric amount. Therefore, the data rate can be determined more accurately.

In the decoding device according to the fourth aspect,
Based on the output of the Viterbi decoding unit 1, thresholds for the number of symbol errors and the path metric amount are set for each data rate that the traffic channel can take, and this threshold is compared with the number of symbol errors and the path metric amount. Then, the data rate of the traffic channel is determined based on the comparison result. Therefore, the data rate can be easily determined.

In the decoding device according to claim 5,
The threshold setting unit 23 or 42 sets a threshold value based on the output of the Viterbi decoding unit 1 corresponding to the sync channel. Therefore, since there is a time margin from setting the threshold value to determining the data rate, it is possible to prevent a heavy load on the device.

In the decoding device according to the sixth aspect,
The threshold setting unit 23 or 42 sets the threshold value based on the output of the Viterbi decoding unit 1 corresponding to the traffic channel. Therefore, since the threshold value is set immediately before the data rate is determined, the data rate can be determined with high accuracy.

In the decoding device according to the seventh aspect,
A threshold value is set for each reliability. Then, the number of symbol errors and the path metric amount are sequentially compared with those having a high reliability and one having a higher data rate, and then the one having a lower reliability by one. The rates are compared in sequence from the highest to the lowest. Therefore, when the number of symbol errors and the amount of path metric correspond to the threshold value of the high reliability and high data rate, the data rate can be immediately determined.

In the decoding device according to claim 8,
The threshold value setting unit 23 sets a threshold value for each reliability. The disconnection standard estimated when the data rate of the traffic channel is determined is the threshold setting unit 2
When 3 is worse than the line disconnection standard estimated when the threshold is set, the reliability of the threshold applied to determine the data rate of the traffic channel is reduced and output. Therefore, the reliability of the determination result of the data rate can be known, and when the reliability is low, the determination result can be prevented from being used, thereby preventing malfunction of the device, for example. become.

In the decoding device according to the ninth aspect,
Based on the error detection result of the CRC error detection unit 3, the order of the thresholds set for each data rate that can be taken by the traffic channel is determined for comparison with the number of symbol errors and the path metric amount. Therefore, the data rate can be determined quickly and more accurately.

In the decoding device according to the tenth aspect, the number of symbol errors and the path metric amount are compared in order from the threshold value corresponding to the data rate that the traffic channel is likely to take. Therefore, the data rate can be determined more quickly.

[0030]

Embodiments of the present invention will be described below with reference to the drawings.

[First Embodiment] FIG. 1 is a block diagram showing the configuration of a first embodiment of a mobile radio such as a cellular telephone to which the decoding device of the present invention is applied. In this mobile wireless device, communication is performed by, for example, the CDMA system.

In the present embodiment, the convolutional code used is, for example, a code having a coding rate of 1/2 and a constraint length of 9, and is composed of the following generator polynomials. ^{G1 (x) = x 8 +} x 7 + x 5 + x 3 + x 2 + x + 1 G2 (x) = x 8 + x 4 + x 3 + x 2 +1

Further, it is assumed that the path memory length of the Viterbi decoding unit 1 described later is 48 for the sync channel and 64 for the traffic channel, for example.

FIG. 2 shows the specifications of each channel received by this mobile radio device. FIG. 2A shows the structure of a sync channel. This sync channel has a frame period of 20 ms for one frame, for example, and is 96
It is adapted to be continuously transmitted at a known fixed rate of 00 bps. Further, FIG. 2B shows a structure of a traffic channel, and this traffic channel has, for example, a frame cycle of 20 ms for one frame.
At 9600bps, 4800bps, 2400bps
One of the four data rates of s and 1200 bps is selected on a frame-by-frame basis for continuous transmission. In FIG. 2B, 4
800 bps, 9600 bps, 2400 bps, 12
This indicates that the data has been transmitted in the order of 00 bps.

Further, FIG. 2C shows the frame structure of the sync channel, and one frame has 192-bit data (for example, for synchronizing the transmitting side and the mobile radio of FIG. 1). It is designed to consist of time information, etc.). Also, FIG. 2D shows traffic channels of 9600 bps, 4800 bps, and 2400 b.
9 shows a frame structure of each of ps and 1200 bps, where 9600 bps is data (for example, voice data) as original information of 172 bits, data for error detection of 12 bits (for example, CRC code), and 8 bits for all 0 tail bits, 4
800 bps is 80-bit data, 8-bit CR
With a C code and all 0 tail bits of 8 bits,
2400 bps is 40 bits of data and 8 bits of all 0 tail bits, and 1200 bps is 1
It is configured to include 6-bit data and 8-bit all-zero tail bits.

In the mobile radio device shown in FIG. 1, two channels are switched in order of the sync channel and the traffic channel for reception, and the convolutionally encoded data is subjected to the maximum likelihood decoding by the Viterbi algorithm to obtain the traffic channel. The data rate judgment is performed.

That is, the Viterbi decoding unit 1 converts an input symbol (modulation symbol) received and demodulated by a reception circuit (not shown) into a known fixed data rate of a sync channel and four types of data rates of a traffic channel. Maximum likelihood decoding is performed on a frame-by-frame basis, for example, based on the Viterbi algorithm. Further, the Viterbi decoding unit 1 outputs the path metric amount of the maximum likelihood path at the time of maximum likelihood decoding.

The hard decision symbol storage unit 2 makes a hard decision on the input symbol and stores the result. When an error detection code such as a CRC code is added to the decoded data decoded by the Viterbi decoding section 1, the CRC error detection section 3 performs an error detection calculation based on this CRC code, and the error is calculated. The detection calculation result is output. CRC storage unit 4
Stores the CRC error detection calculation result (CRC calculation result) output from the CRC error detection unit 3. The convolutional encoder 5 encodes the decoded data decoded by the Viterbi decoding unit 1 with the same coding rate 1/2 and constraint length 9 as the transmitting side, and outputs the result (symbol).

The symbol error number calculation unit 6 compares the hard decision symbols determined by the hard decision symbol storage unit 2 with the encoded symbols encoded by the convolutional encoder 5 in symbol units. Then, for the sync channel, a known fixed data rate (for example, 9600 bp)
s, 4800 bps, and 2400 bps, and only one that is known in advance (9600 bps as described above in the present embodiment), four types of data for the traffic channel. Rate (9600bps, 4800bps,
2400 bps, 1200 bps)
When there is an error in the compared symbols (when the hard decision symbol and the encoded symbol do not match), the count value of the built-in counter is incremented by 1, and when there is no error (the hard decision symbol and the encoded symbol match). In this case), the process of not changing the count value is performed for each frame, and the result is output.

The symbol error number storage unit 7 stores the number of symbol errors calculated by the symbol error number calculation unit 6. The path metric amount storage unit 8 obtains the path metric amount at the time of decoding the input final symbol for each frame, which is obtained by the Viterbi decoding unit 1, for each of the known fixed data rate of the sync channel and each of the four types of data rates of the traffic channel. Remember. The decoded data storage unit 9 outputs the decoded data, which is output from the Viterbi decoding unit 1 and which is obtained by decoding the sync channel at a known fixed data rate and the decoded data obtained by decoding the traffic channel at each of four data rates, to a frame. Remember in units.

The data rate determination unit 10 determines, for each frame, the data rate that has the highest probability of being transmitted among all the data rates that can be transmitted by the traffic channel. The channel selector 11 is the Viterbi decoding unit 1
And the control signal for causing the data rate determination unit 10 to perform the processing corresponding to each of the sync channel and the traffic channel is output.

Next, FIG. 3 shows a detailed configuration of the data rate determination unit 10 of FIG. In this figure, the normalization calculation unit 21 stores 9600 bps, 4800 bps, and 2400 b stored in the symbol error number storage unit 7 of FIG.
The number of symbol errors for ps and 1200 bps is normalized, and the result (hereinafter, referred to as the number of normalized symbol errors) is output. The average value calculation unit 22 stores 9600 bps, 480 stored in the path metric amount storage unit 8 of FIG.
The average value of each effective path metric amount is calculated from the path metric amount for each frame of 0 bps, 2400 bps, and 1200 bps, and the result (hereinafter, referred to as average value of path metric amount) is output.

The threshold setting unit 23, as will be described later,
Four types of data rates (960
0 bps, 4800 bps, 2400 bps, 1200
It stores a table of threshold values (hereinafter referred to as a threshold table) classified according to Eb / N0 (energy / noise in bit units) (line disconnection standard) for determining bps). Then, corresponding to Eb / N0 output from the Eb / N0 estimation unit 26, one of the stored threshold tables is output to the threshold storage unit 27 and stored (in the threshold setting storage unit 27. Set). Threshold storage unit 27
Stores a threshold table output from the threshold setting unit 23.

The comparison section 24 calculates the CRC calculation result stored in the CRC storage section 4 of FIG. 1, the normalized symbol error number output from the normalization calculation section 21, and the path metric amount output from the average value calculation section 22. Each average value is compared with each threshold value of the threshold value table stored in the threshold value storage unit 27, and the comparison result is output to the determination unit 25. The determination unit 25
The data rate is determined from the comparison result output from the comparison unit 24, and the determination result and the reliability indicating the degree of reliability of the determination result (in this embodiment, the smaller the value, the higher the reliability. ) Is output.

The Eb / N0 estimation unit 26 is a normalization calculation unit 2
Normalized symbol error count from 1 and average value calculation unit 2
Eb / N0 is estimated from the average value of the path metric amount from 2 based on the experimental result described later.

4 to 7, the traffic channel data rates are 9600 bps and 4800 b.
When decoding is performed assuming that ps, 2400 bps, and 1200 bps are fixed, in each case, 9600 bps,
4800bps, 2400bps, and 1200bp
It is a figure which shows the average value of the path metric amount obtained from the experimental data of s, and the average value of the number of normalized symbol errors.

In creating these figures, Eb / N0 was used as data for experiments, using 10,000 frames and WGN (white Gaussian noise) channel.
= + 1, +2, +3, +4 (dB), the average value of the path metric amount and the average value of the normalized symbol error numbers were obtained.

From the figure, when the data rate is correct (the data rate of the experimental data is equal to the assumed data rate), when there is no noise, both the path metric amount average value and the normalized symbol error number average value are obtained. It can be seen that the value becomes 0, and the value increases as Eb / N0 worsens.

When the data rate is incorrect (the data rate of the experimental data is not equal to the assumed data rate), the average value of the path metric amount and the normalized symbol are irrespective of the value of Eb / N0. It can be seen that the average number of errors shows a substantially constant value.

In the above-mentioned threshold setting unit 23, based on the experimental results shown in FIGS. 4 to 7, the data rates of the traffic channels are 9600 bps and 4800 bp.
A threshold table in which thresholds of the path metric amount average value and the normalized symbol error number that can be regarded as s, 2400 bps, and 1200 bps are described is stored. The threshold table stored in the threshold setting unit 23 is, as shown in Table 1 and Table 2 described later,
The CRC calculation result by the CRC error detector 3 is taken into consideration.

Further, the Eb / N0 estimation unit 26 stores the experimental results shown in FIGS. 4 to 7 (portions shown by dotted lines in the drawings).

For each of the four data rates of the traffic channel, the average value calculation unit 22
The decoded data (bit position of the decoded data) required to calculate the average value of the path metric amount per frame is
In FIG. 8, it is a portion indicated by hatching.

Next, a fixed data rate of the sync channel (for example, 9600 bps) is received, and then the sync channel is switched to the traffic channel,
The operation of receiving a traffic channel and determining its data rate will be described.

When decoding the sync channel, first, as described above, the sync channel is received by the receiving circuit. When the demodulation is successful, the receiving circuit outputs to the channel selector 11 a channel control signal informing that the sync channel is being received and the data rate. Then, the channel selector 11 instructs the Viterbi decoding unit 1 and the data rate determination unit 10 that a sync channel having a fixed data rate is input.

The Viterbi decoding unit 1 continuously applies the modulation symbols of the sync channel input for each frame to a fixed data rate (as described above, 9600).
bps), the decoded data is output to the convolutional encoder 5 and the decoded data storage unit 9, and the path metric amount of the maximum likelihood path is output to and stored in the path metric amount storage unit 8 for each 1-bit decoding.

The hard decision symbol storage unit 2 makes a hard decision of the received symbol from the modulation symbols input at the same timing as the Viterbi decoding unit 1, and stores the result. The convolutional encoder 5 re-encodes the decoded data input from the Viterbi decoding unit 1 and outputs the encoded symbol to the symbol error number calculation unit 6.

In the symbol error number calculation unit 6, the hard-decision symbol stored in the hard-decision symbol storage unit 2 is read at the timing when the convolutional encoder 5 outputs the encoded symbol corresponding to the hard-decision symbol. Then, the exclusive OR of the hard decision symbol and the encoded symbol is calculated. Then, if the result is 1, the count value of the built-in counter is incremented by 1, and if the result is 0, the operation of keeping the count value for one frame is performed, and the result is stored in the symbol error number storage unit 7. Output and store. After that, the symbol error number calculation unit 6
Resets the count value of the counter and repeats the above operation.

The symbol error number stored in the symbol error number storage unit 7 and the path metric amount stored in the path metric amount storage unit 8 are input to the data rate determination unit 10.

In the data rate determination unit 10 (FIG. 3), the normalization calculation unit 21 calculates the number of normalized symbol errors from the number of symbol errors, and the average value calculation unit 22 calculates the path for one frame from the path metric amount. The metric amount average value is calculated and output to the Eb / N0 estimation unit 26.

At the same time, the judging section 25 outputs the decoded data at the data rate (in this case, 9600 bps) input from the channel selector 11 (in this case, 9600 bps) to the decoded data storage section 9 (FIG. 1). Output the data rate control signal. As a result, the decoded data storage unit 9 temporarily stores the decoded data of the sync channel output from the Viterbi decoding unit 1 and then the data rate determination unit 1
The data is output in frame units at a data rate (9600 bps) according to the data rate control signal from 0.

Further, the Eb / N0 estimation unit 26 determines a fixed data rate of the sync channel (960 as described above) from the normalized symbol error number and the path metric average value.
Experimental result corresponding to 0 bps (in the present embodiment,
E based on the average value of the path metric amount at 9600 bps shown in FIG.
b / N0 is estimated and output to the threshold setting unit 23.

In the threshold setting unit 23, the Eb / N0 estimation unit 2
The threshold value table corresponding to Eb / N0 output from No. 6 is output to and stored in the threshold value storage unit 27.

The above operation is performed during reception of the sync channel.

Here, the data rate of the sync channel is set to 9600 bps as described above, and the normalization symbol output from the normalization calculation unit 21 or the average value calculation unit 22 immediately before the transition from the sync channel to the traffic channel, for example. Assuming that the number of errors or the average value of the path metric amount is 2 or 1500, respectively, the estimated Eb / N0 output from the Eb / N0 estimation unit 26 is +2 dB from FIG. In this case, the threshold value setting unit 23 outputs the threshold value tables as shown in Tables 1 and 2 to the threshold value storage unit 27 and stores them.

[0065]

[Table 1]

[0066]

[Table 2]

After that, when the sync channel is switched to the traffic channel and the demodulation of the traffic channel is successful, a channel control signal notifying that the traffic channel is being received is input from the receiving circuit to the channel selector 11. The channel selector 11 notifies the Viterbi decoding unit 1 and the data rate determination unit 10 that the traffic channel will be processed.

Then, the Viterbi decoding unit 1 stores the modulation symbol input for each frame in the input buffer having the built-in modulation symbol, and the data rate thereof is, for example, 9600b.
ps, 4800bps, 2400bps, 1200bp
Decode assuming the order of s. Then, the decoded data obtained as a result is output to the CRC error detection unit 3, the convolutional encoder 5, and the decoded data storage unit 9, and
For each bit decoding, the path metric amount of the maximum likelihood path is output and stored in the path metric amount storage unit 8 in the order of the decoded data rate.

The hard-decision-symbol storage unit 2 determines the hard-decision of the received symbol from the modulation symbol input at the same timing as the Viterbi decoding unit 1, for example, 9600 bp.
s, 4800bps, 2400bps, 1200bps
And the result is stored. Convolutional encoder 5
From the Viterbi decoding unit 1 to 9600 bps, 4800 b
The decoded data output assuming the data rate in the order of ps, 2400 bps, and 1200 bps is re-encoded while resetting the internal state for each data rate, and the encoded symbol is output to the symbol error number calculation unit 6.

The received symbol stored in the hard-decision symbol storage unit 2 and the encoded symbol from the convolutional encoder 5 corresponding to the received symbol are input to the symbol error number calculation unit 6 at the same timing. 960
0 bps, 4800 bps, 2400 bps, 1200
The exclusive OR of the received symbol and the encoded symbol is taken in the order of bps. Then, the symbol error number calculation unit 6 increments the count value of the built-in counter by 1 if the result is 1, and leaves the count value as it is if 0, at 9600 bps, 48
The counter is reset every time the data rate changes in the order of 00 bps, 2400 bps, and 1200 bps, each time the data rate changes, the result is output to the symbol error number storage unit 7 for each data rate, and then the counter is reset.

The CRC error detection unit 3 outputs the decoded data of the data rate output from the Viterbi decoding unit 1 to which the CRC code is added, that is, 9600 bps and 4800 bps.
CRC is calculated for the decoded data of, and the calculation result is set to 0 when no error is detected for each data rate, and the calculation result is set to 1 when an error is detected and output to the CRC storage unit 4. To do. CRC storage unit 4
Then, the calculation result (CRC calculation result) from the CRC error detection unit 3 is temporarily stored.

The data rate determination unit 10 (FIG. 3) has a C
9600 bps, 480 stored in the RC storage unit 4
CRC calculation results of 0 bps, 9600 bps and 4800 bp stored in the symbol error number storage unit 7, respectively.
s, 2400 bps, 1200 bps, respectively, the number of symbol errors, 9600 bps, 4800 bps, 2400 bps, 1 stored in the path metric amount storage unit 8.
The path metric amount of 200 bps is input.

The CRC calculation result is directly input to the comparison section 24. After the symbol error number is normalized by the normalization calculation unit 21 to be the normalized symbol error number, and the path metric amount is averaged by the averaging calculation unit 22 to be the path metric amount average value. , And is output to the comparison unit 24 and the Eb / N0 estimation unit 26.

As shown in the flowchart of FIG. 9, the comparison unit 24 calculates the CRC calculation results of 9600 bps and 4800 bps, respectively, 9600 bps, 4800 bps, 2
Normalized symbol error numbers and path metric amount average values at 400 bps and 1200 bps, respectively, and Table 1 and Table 2.
From the threshold value shown in Table 1 for the determination data rate of 9600 bps with the reliability level 1 shown in Table 1, the reliability level 2 shown in Table 2
The determination data rate is compared in the order of the threshold value of 1200 bps, and the reliability and the determination data rate for which the comparison results are all true are output to the determination unit 25 and the Eb / N0 estimation unit 26.

That is, for example, the comparison unit 24 has 9600 bp.
CRC calculation results of s and 4800 bps are 0 and 1, respectively.
Is 9600 bps, 4800 bps, 240
The normalized symbol error numbers of 0 bps and 1200 bps are
70 or less, 80 to 110 range, 80 to 1 respectively
A value in the range of 10 and 80 to 110,
The value of 00 bps is the minimum value, and 9600
bps, 4800bps, 2400bps, 1200b
The path metric amount average value of ps is 1800 or less, in the range of 3200 to 3800, and 3000 to 330, respectively.
A value in the range 0, 1800 to 2200,
When the value of 9600 bps is the minimum value (when the conditions at the top of Table 1 are all true), the data rate is set to 9
The determination is made as 600 bps, and the reliability with a value of 1 is output.

Further, for example, the comparison unit 24 is 9600 bp.
s, 4800 bps CRC calculation result is 1, 1
Is 9600 bps, 4800 bps, 240
The normalized symbol error numbers of 0 bps and 1200 bps are
The values are 80 to 120, 70 to 110, 70 to 110, and 80 or less, and are 9600 bps, 4800 bps, and 2400 bp.
The average value of the path metric amounts of s and 1200 bps is in the range of 1800 to 2500 and 2300 to 350, respectively.
0 range, 1800 to 3000 range, values less than 1800 and 1200 bps is the minimum value (when all the conditions at the bottom of Table 2 are true), the data rate is determined to be 1200 bps At the same time, the reliability value of 2 is output.

The comparing section 24 outputs a data rate determination inability signal indicating that the data rate determination could not be performed if none of the comparison results are true.

On the other hand, the Eb / N0 estimation unit 26 is connected to the comparison unit 2
The determination data rate output from 4 is 9600 bps,
In the case of 4800 bps, 2400 bps, and 1200 bps, Eb / N0 is estimated from the normalized symbol error number and the path metric average value by referring to the dotted line parts shown in FIGS. Judgment unit 25
Output to.

The judgment unit 25 outputs the reliability output from the comparison unit 24 and the E output from the Eb / N0 estimation unit 26.
The final reliability is determined (determined) from b / N0 and output to the decoded data storage unit 9 together with the determination data rate from the comparison unit 24.

In the judging section 25, when the value of Eb / N0 estimated by the Eb / N0 estimating section 26 is worse than, for example, +2 dB, when the reliability from the comparing section 24 is 1, it becomes 2
When it is 2, it is output as 3 and is also output as Eb / N0.
When the value of is better than +2 dB, the reliability of the comparison unit 24 is output as it is.

The data rate (determination data rate) output from the determination section 25 and its reliability are supplied to the decoded data storage section 9 (FIG. 1). In the decrypted data storage unit 9,
9600 bps, 48 output from the Viterbi decoding unit 1
The data rate determination unit 10 (determination unit 25) of the decoded data for one frame decoded assuming the data rates of 00 bps, 2400 bps, and 1200 bps.
The decoded data decoded on the assumption that the data rate is the same as the determination data rate supplied from the device is output as accurate decoded data for one frame. Further, the decoded data storage unit 9 outputs the reliability of the data rate supplied from the data rate determination unit 10.

When the data rate determination unit 10 outputs the data rate determination impossible signal, the decoded data storage unit 9 does not output the decoded data but outputs a signal indicating the determination is impossible.

Here, the reliability of the data rate determination result can be used when it is desired to reduce the erroneous determination of the data rate even if the number of cases in which the determination cannot be made increases. That is, for example, when the reliability is 3, the decoded data storage unit 9 is not caused to output the decoded data, but the signal indicating that the determination is impossible is output, thereby reducing the erroneous determination of the data rate. it can.

Thus, according to the above apparatus, the data rate for decoding can be accurately and satisfactorily determined by determining the data rate for each unknown frame of the traffic channel with a simple configuration.

That is, the mobile wireless device shown in FIG.
Like the traffic channel of the A system,
For each frame on the same channel, decoding of symbols convolutionally coded by the same convolutional code and having different data rates and determination of the data rate are performed, and decoding of the variable data rate for outputting decoded data in frame units is performed. ) CRC calculation result (2) Symbol error number (3) Path metric amount Three parameters are used as parameters for data rate determination and reliability determination.

Then, decoding and calculation of (1) to (3) are performed frame by frame for all data rates that may have been transmitted, and all decoded data are stored, and (1) to (3) are stored. From the calculation result of (), normalization is performed for (2) and averaging is performed for (3), and comparison is made with a threshold value corresponding to a preset data rate and reliability.

In this case, the reliability is high, and the comparison is performed in the order from the threshold for the fast data rate to the threshold for the slow data rate, and then the reliability is lowered by one to reduce the threshold from the fast data rate to the slow data rate. Compare in the order of. Then, if there is a threshold that applies, the comparison is terminated by using the data rate and the reliability corresponding to the threshold as arguments, and the reliability is estimated from the number of symbol errors and the path metric amount of the determined data rate. When Eb / N0 is worse than Eb / N0 at the time of setting the threshold value, the reliability is lowered by one rank, the data rate determination is ended, and the decoded data of the determined reliability and the determined data rate is output. Do this for each frame.

As described above, since the data rate judgment and reliability are obtained from the CRC, the number of symbol errors, and the path metric amount for all the data rates that may have been transmitted, the erroneous judgment is made. The probability can be reduced. The threshold is set based on Eb / N0 estimated from the number of symbol errors and the amount of path metric of the sync channel immediately before the traffic channel transition. Can be done by

Further, the final reliability is determined by Eb / N0 estimated from the number of symbol errors of the determined data rate and the path metric amount. Then, when the Eb / N0 estimated from the number of symbol errors of the determined data rate and the path metric amount is worse than the set threshold Eb / N0, the reliability is set to be low. By taking these into consideration, it becomes possible to increase the reliability of reliability.

In addition, when the comparison is performed in the order of high reliability and a threshold for a fast data rate to a threshold for a slow data rate, then the reliability is lowered by one and the comparison is performed in the same manner. If the current Eb / N0 is a value equal to or higher than the Eb / N0 when the threshold is set, the data rate corresponding to the threshold and the reliability are terminated as arguments. It becomes unnecessary to compare the cases, and the load on the device can be reduced.

Further, in the calculation of the average value of the path metric amount, as shown in FIG. 8, the average of only the path metric amount of the effective maximum likelihood path of the decoded data burst-decoded in frame units is calculated. Since this is set to a more probable value, it is possible to reduce the erroneous determination rate of the data rate.

Further, in the above apparatus, the specifications of the sync channel and the traffic channel are as shown in Table 3.

[0093]

[Table 3]

In Table 3, as the modulation symbol, when the data rate is 9600 bps, the convolutionally encoded received symbol itself is used, and 480
In the case of 0 bps, the value of the convolutionally encoded symbol is used twice, and in the case of 2400 bps, the value of the convolutionally encoded symbol is used four times,
In the case of 1200 bps, the value of the convolutionally coded symbol is used repeatedly eight times, so that the input buffer incorporated in the Viterbi decoding unit 1 uses the data when decoding one frame of data. Regardless of the rate, 384 modulation symbols (for one frame) may be stored.

The puncture symbols are generated only in the traffic channel, every 24 modulation symbols, and in the continuous 2 modulation symbols therein. Then, in the received modulation symbol, the puncture symbol is determined by a circuit (not shown) in the preceding stage of the Viterbi decoding unit 1, and the modulation symbol determined as the puncture symbol is set to 0 without 6-bit soft decision. Has been done. Therefore, the extraction of the puncture symbol may be performed when the value of the received symbol is 0.

Further, the calculation of the number of normalized symbol errors in the normalization calculation unit 21 (FIG. 3) is performed as follows (however, the sync channel is, for example, 9600 bps, 4800 bps, or 2400 bps). Only about).

Number of normalized symbol errors of 9600 bps = (number of symbol errors of 9600 bps calculated by the symbol error number calculation unit 6) × 1 Number of normalized symbol errors of 4800 bps = (4800 bps calculated by the number of symbol error calculation unit 6 Symbol error number) × 2 2400 bps normalized symbol error number = (2400 bps symbol error number calculated by the symbol error number calculation unit 6) × 4 1200 bps normalized symbol error number = (in the symbol error number calculation unit 6 Calculated 1200 bps symbol error count) × 8 ... (1)

Further, the calculation of the average value of the path metric amount in the average value calculation unit 22 (FIG. 3) is performed as follows by using the hatched portion in FIG. 8 ( However, when the path memory length built in the Viterbi decoding unit 1 is 64).

In the case of 9600 bps (the number of data bits in one frame> path memory length) average value of path metric = (path metric of maximum likelihood path corresponding to 128 (= 192-64) bits of first decoded data of frame) Total amount) / 12
In the case of 4 · 4800 bps (the number of data bits in one frame> path memory length), the average value of the path metric amount = (the path metric amount of the maximum likelihood path corresponding to 32 (= 96−64) bits of the first decoded data of the frame) Total) / 32 ・ In the case of 2400 bps (the number of data bits in one frame <path memory length) average value of path metric amount = (path metric amount of maximum likelihood path corresponding to 1 bit of the first decoded data in the frame) ・ 1200 bps Case (number of data bits in one frame <path memory length) Average value of path metric amount = (path metric amount of maximum likelihood path corresponding to 1 bit of first decoded data in frame) (2)

[Second Embodiment] By the way, in the above first embodiment, the calculation formula of the normalized symbol error number in the normalization calculation unit 21 (FIG. 3) shown in the formula (1) is a puncture symbol. This is not taken into consideration. Therefore, it is conceivable that an error occurs in the number of normalized symbol errors, which causes the determination rate of the data rate to deteriorate somewhat.

Further, for example, when an error that cannot detect an error occurs in the data, or by chance, the CRC calculation result of the data of 9600 bps is actually 48.
The CRC calculation result of 00 bps is equal to or the CRC calculation result of 4800 bps data is actually 9
It can be considered to be equal to the CRC calculation result of 600 bps. In addition, the wrong CRC calculation result, or CR
The C calculation result happens to be 9600 bps and 4800 b
It is also conceivable that both CRC calculation results of ps are equal.

When the CRC calculation result is erroneous as described above, the correct data rate must be determined in the threshold table that absolutely views the CRC calculation results shown in Tables 1 and 2. Can not be.

Therefore, FIG. 10 is a block diagram showing the configuration of a second embodiment of a mobile radio to which the decoding device of the present invention is applied. In addition, in the figure, the same reference numerals are given to the portions corresponding to the case in FIG.

Therefore, this mobile wireless device is provided with a data rate determining unit 32 in place of the data rate determining unit 10 and additionally has a new puncture symbol detecting unit 3.
1 is provided and is configured similarly to the mobile wireless device shown in FIG.

The puncture symbol detector 31 detects a puncture symbol (which is set to 0 without 6-bit soft decision as described above) from the input modulation symbols. Then, the symbol error number calculation unit 6 is controlled to exclude the symbol corresponding to the detected puncture symbol from the target of the symbol error count. Therefore, in the symbol error number calculation unit 6,
Symbol errors are counted by excluding the symbols corresponding to the puncture symbols detected by the puncture symbol detection unit 31.

The data rate determining unit 32 is constructed as shown in FIG. In the figure, the same reference numerals are given to the portions corresponding to those in FIG. Therefore, the data rate determination unit 32 is equivalent to the normalization calculation unit 2
1. Instead of the threshold value setting unit 23 or the comparison unit 24, a normalization calculation unit 43, a threshold value setting unit 42, or a comparison determination unit 41.
Are provided respectively, and the determination unit 25 is deleted, and the configuration is similar to that of the data rate determination unit 10 in FIG.

In the normalization calculation unit 43, 9600 bps, 480 stored in the symbol error number storage unit 7 of FIG.
From the symbol error numbers of 0 bps, 2400 bps, and 1200 bps, the respective normalized symbol error numbers are calculated according to the following equation in consideration of the puncture symbol.

Number of normalized symbol errors of 9600 bps = (number of symbol errors of 9600 bps calculated by the symbol error number calculation unit 6) × 1 × (11/12) × (172/192) Number of normalized symbol errors of 4800 bps = (Number of symbol errors of 4800 bps calculated by the symbol error number calculation unit 6) × 2 × (11/12) × (80/96) Number of normalized symbol errors of 2400 bps = (Calculated by the symbol error number calculation unit 6 Number of symbol errors of 2400 bps) × 4 × (11/12) × (40/48) Number of normalized symbol errors of 1200 bps = (Number of symbol errors of 1200 bps calculated by the symbol error number calculator 6) × 8 × (11 /12)×(16/24)...(3)

In the equation (3), 11/12 (=
(24-2) / 24) is a coefficient for correcting the influence of the puncture symbol of 2 symbols of 24 symbols mentioned above. Also, 172 / 192,80 / 9
6,40 / 48, 16/24 each is 9600bp
s, 4800bps, 2400bps, 1200bps
Of the CRC code or the tail bit of the data of 1 frame of (1) is converted (the number of symbol errors calculated by the symbol error number calculation unit 6 is converted into the number of symbol errors of bits representing substantial information). Is a coefficient for.

At the time of receiving the sync channel, as in the case of the first embodiment, the normalized symbol error number calculated by the normalization calculation unit 43 is the path metric amount average calculated by the average value calculation unit 22. Along with the value, Eb / N
0 is output to the estimation unit 26. Then, in the Eb / N0 estimation unit 26, based on one of the experimental results shown in FIGS. 4 to 7 corresponding to the fixed data rate of the sync channel, from the normalized symbol error number and the path metric average value, Eb / N0 is estimated and output to the threshold setting unit 42.

In the threshold value setting section 42, the data rates of the traffic channels are 9600 bps and 4800 bp based on the experimental results shown in FIGS. 4 to 7.
A threshold table in which thresholds of a path metric amount average value and a normalized symbol error number that can be regarded as s, 2400 bps, and 1200 bps are described,
Those classified according to Eb / N0 are stored. The threshold value table stored in the threshold value setting unit 42 is different from the threshold value table of the threshold value setting unit 23 in the first embodiment and does not consider the CRC calculation result by the CRC error detection unit 3. .

The threshold value setting unit 42 includes the Eb / N0 estimation unit 26.
When Eb / N0 is received from, the threshold table corresponding to the Eb / N0 is output to and stored in the threshold storage unit 27 (set in the threshold storage unit 27).

The above operation is performed during reception of the sync channel.

Table 4 shows an example of the threshold value table set in the threshold value storage unit 27 by the threshold value setting unit 42 when the Eb / N0 output from the Eb / N0 estimation unit 26 is +3 dB.

[0115]

[Table 4]

After that, when the sync channel is switched to the traffic channel and the demodulation of the traffic channel is successful, the Viterbi decoding unit 1 (FIG. 10) inputs the traffic channel for each frame in the same manner as in the first embodiment. Decoding the modulation symbols of the data rate of 9600 bps, 4
Assuming that the order of 800 bps, 2400 bps, and 1200 bps is started, the data rate determination unit 32 thereby stores 9600 b stored in the CRC storage unit 4.
As a result of each CRC calculation of ps and 4800 bps, 960 counted excluding the symbol corresponding to the puncture symbol stored in the symbol error number storage unit 7
0 bps, 4800 bps, 2400 bps, 1200
Number of symbol errors for each bps, 9600 bps, 4800 bp stored in the path metric amount storage unit 8
The path metric amounts of s, 2400 bps, and 1200 bps are input.

In the data rate judging section 32 (FIG. 11),
The CRC calculation result is directly input to the comparison / determination unit 41. Then, the number of symbol errors is normalized by the normalization calculation unit 43 according to the equation (3) to be a normalized number of symbol errors, and the path metric amount is averaged by the averaging calculation unit 22 to obtain the path metric amount. The average value is output to the comparison / determination unit 41.

The comparison / determination unit 41 performs the processing according to the flowchart shown in FIG. That is, first, in step S1, the CRC calculation result (CRC detection result) is determined, and when it is determined that it is correct only at 9600 bps, the rate determination process of step S2 is performed.

When it is determined in step S1 that the CRC calculation result is correct only at 4800 bps,
The rate determination process of step S3 is performed, and 9600 bp
When it is determined that both s and 4800 bps are correct, the rate determination process of step S4 is performed. Further, in step S1, the CRC calculation result is 9600.
When it is determined that both bps and 4800 bps are incorrect, the rate determination process of step S5 is performed.

In the rate judgment processing of steps S2 to S5, 96 steps are performed in the order based on the CRC calculation result and the probability of occurrence of the data rate that the traffic channel can take.
00 bps, 4800 bps, 2400 bps, 120
The normalized symbol error number and path metric amount average value for each 0 bps are compared with the threshold value (threshold value set) of the threshold value table shown in Table 4, and the reliability and the determination data rate (determination value) for which all comparison results are true. The target data rate) is output.

The data rates of the traffic channel are 9600 bps, 4800 bps, 240
The occurrence probability of 0 bps and 1200 bps is 3 /
Experiments have shown that it is around 8 weak, 1/8 weak, 1/8 strong, and 3/8 strong. Therefore, the probability of occurrence of the data rate of the traffic channel is 1200b.
ps, 9600bps, 2400bps, 4800bp
It becomes higher in the order of s.

That is, in the rate determination processing of step S2, first, in step S11 shown in FIG. 13, the CRC calculation result of the threshold value set of the threshold value table shown in Table 4 stored in the threshold value storage unit 27 (FIG. 11) is calculated. Is correct 96
A threshold set of 00 bps, which has a high reliability, and a threshold of threshold set 1 and four data rates (96
00 bps, 4800 bps, 2400 bps, 120
0 bps) The respective normalized symbol error numbers and the average value of the path metric amount are compared.

In step S11, when the comparison results of the thresholds of the threshold set 1, the normalized symbol error numbers of the four data rates, and the average value of the path metric amount are all true, that is, 9600 bps, 4800 bps,
Of the normalized symbol error numbers of 2400 bps and 1200 bps, 9600 bps is the smallest and smaller than 40, and further 9600 bps and 4800 bps.
When the path metric amount average value of s, 2400 bps, and 1200 bps is the smallest one of 9600 bps, the process proceeds to step S12, and the data rate is 96.
The reliability is determined to be 1 at 00 bps, and the process ends.

If any of the comparison results of the threshold of the threshold set 1 in step S11, the number of normalized symbol errors for each of the four data rates, and the average value of the path metric amount is not true, Proceeding to S13, of the threshold value sets of the threshold value table shown in Table 4 stored in the threshold value storage unit 27 (FIG. 11), the threshold value set of the correct CRC calculation result is 9600 bps and has the second highest reliability. The threshold value of a certain threshold value set 2 is compared with the normalized symbol error number and the path metric amount average value for each of the four data rates.

If the comparison results of the thresholds of the threshold set 2 in step S13 with the normalized symbol error numbers of the four data rates and the average value of the path metric amount are all true, the process proceeds to step S14, and the data rate is changed. The reliability is determined to be 2 at 9600 bps, and the processing ends.

If any of the comparison results of the threshold of the threshold set 2 in step S13, the number of normalized symbol errors for each of the four data rates, and the average value of the path metric amount is not true, that is, shown in Table 4. In the threshold set of the threshold table, neither the condition of the threshold set 1 or 2 of 9600 bps that the CRC calculation result is the correct data rate is satisfied with the normalized symbol error number and the path metric amount average value of each of the four data rates. In the following steps S15, S17, and S19, the threshold set, the number of normalized symbol errors for each of the four data rates, and the path metric amount average value are arranged in the order based on the probability of occurrence of the data rate that the traffic channel can take. And are compared.

Therefore, first in step S15,
The number of normalized symbol errors for each of the four data rates,
The average value of the path metric amount and the highest occurrence probability of 120
The 0 bps threshold set 5 is compared. Then, if the comparison results of the normalized number of symbol errors and the average value of the path metric amount for each of the four data rates and the threshold value of the threshold value set 5 are all true, the process proceeds to step S16, and the reliability is set at the data rate of 1200 bps. Is determined to be 1, and the process ends.

If any of the comparison results of the normalized symbol error number and the path metric amount average value for each of the four data rates and the threshold value of the threshold value set 5 is not true, the process proceeds to step S17. Threshold set 4 of 2400 bps having the next highest occurrence probability of 1200 bps (note that the highest occurrence probability of 1200 bps is 9600 bps as described above, but in this case, threshold sets 1 and 2 of 9600 bps have already been generated. Since the comparison with the normalized number of symbol errors for each of the four data rates and the average value of the path metric amount has been completed, the threshold value set 4 of 2400 bps having the next highest occurrence probability is used). If the comparison results are all true, the process proceeds to step S18, and the data rate is 2
The reliability is determined to be 1 at 400 bps, and the process ends.

On the other hand, if any one of the comparison results of the normalized symbol error number and the path metric amount average value for each of the four data rates and the threshold value of the threshold value set 4 is not true, the process proceeds to step S19. A comparison is made with the threshold set 3 of 4800 bps, which has the next highest occurrence probability of 2400 bps. Then, when the comparison results of the normalized symbol error number and the path metric amount average value for each of the four data rates and the threshold value of the threshold value set 3 are all true,
In step S20, the data rate is 4800 bps
Then, the reliability is determined to be 1, and the process ends.

If any of the comparison results is not true, the process proceeds to step S21, the determination of the data rate is disabled, and the process ends.

Returning to FIG. 12, in the rate determination processing in step S3, first in step S31 shown in FIG.
Among the threshold value sets of the threshold value table shown in Table 4 stored in the threshold value storage unit 27 (FIG. 11), the threshold value of the threshold value set 3 of 4800 bps with a correct CRC calculation result and the four data rates (9600 bps, 4800 bps, 2400).
bps, 1200 bps) and the number of normalized symbol errors and the average value of the path metric amount are compared.

If the comparison results of the thresholds of the threshold set 3 with the normalized symbol error numbers and the path metric amount average values of the four data rates are all true, step S
In step 32, the data rate is 4800 bps, the reliability is determined to be 1, and the process ends.

If any of the comparison results of the threshold of the threshold set 3 in step S31, the number of normalized symbol errors for each of the four data rates, and the average value of the path metric amount is not true, that is, Of the threshold value sets in the threshold value table shown in Table 4, the condition of the threshold value set 3 of 4800 bps at which the CRC calculation result is a correct data rate is not satisfied by the normalized symbol error number and the path metric amount average value for each of the four data rates. In the following steps S33, S35, S37, and S39, the thresholds are set in the order based on the probability of occurrence of the data rate that the traffic channel can take, and in the case of multiple reliability levels, in the order of high reliability. And the normalized symbol error number and path metric amount average value for each of the four data rates It is compare.

Therefore, first, in step S33,
The number of normalized symbol errors for each of the four data rates,
The average value of the path metric amount and the highest occurrence probability of 120
The 0 bps threshold set 5 is compared. Then, if the comparison results of the normalized number of symbol errors and the average value of the path metric amount for each of the four data rates and the threshold value of the threshold value set 5 are all true, the process proceeds to step S34, and the reliability is obtained at the data rate of 1200 bps. Is determined to be 1, and the process ends.

If any of the comparison results is not true, the process proceeds to step S35 and 1200 bp.
The threshold value set of 9600 bps having the second highest occurrence probability after s and the threshold value set 1 having the highest reliability are compared. Then, if the comparison results of the normalized number of symbol errors and the average value of the path metric amount for each of the four data rates and the threshold value of the threshold set 1 are all true, the process proceeds to step S36, and the reliability is obtained at the data rate of 9600 bps. Is determined to be 1, and the process ends.

If any of the comparison results of the number of normalized symbol errors, the average value of the path metric amount, and the threshold value of the threshold value set 1 at each of the four data rates in step S35 is not true, the process proceeds to step S37. Go ahead,
The threshold value set of 9600 bps is compared with the threshold value set 2 having the next highest reliability. If the comparison results of the normalized number of symbol errors and the average value of the path metric amount for each of the four data rates and the thresholds of the threshold set 2 are all true, the process proceeds to step S38, and the reliability of the data rate is 9600 bps. Is determined to be 2, and the process ends.

If any of the comparison results of the number of normalized symbol errors, the average value of the path metric amount, and the threshold value of the threshold value set 2 at each of the four data rates in step S37 is not true, the process proceeds to step S39. Go ahead,
240, which has the highest probability of occurrence next to 9600 bps with a reliability of 2
A comparison with the 0 bps threshold set 4 is made. And
The number of normalized symbol errors for each of the four data rates,
If all the comparison results between the average value of the path metric amount and the threshold value of the threshold value set 4 are true, the process proceeds to step S40,
When the data rate is 2400 bps, the reliability is determined to be 1, and the process ends.

If any of the comparison results is not true, the process proceeds to step S41, the data rate determination is disabled, and the process ends.

As described above, the data rate with the next highest occurrence rate of 2400 bp is 4800 bp.
However, in this case, since the CRC calculation result of 4800 bps was correct, the threshold set 3 of 4800 bps was set.
And the normalized number of symbol errors for each of the four data rates and the average value of the path metric amount are performed in the first step 31, so the processing is terminated after the comparison processing using the threshold set 4 of 2400 bps. It is designed to do.

Next, in the rate determination processing in step S4 (FIG. 12), first, in step S51 shown in FIG. 15, the threshold value set in the threshold value table shown in Table 4 stored in the threshold value storage unit 27 (FIG. 11) is set. , The threshold value of threshold value set 1 which is a threshold value set of 9600 bps having a high occurrence probability among the correct 9600 bps and 4800 bps of which the CRC calculation result is correct and which has high reliability, and the number of normalized symbol errors for each of the four data rates. , And the average value of the path metric amount is compared.

If all the comparison results are true, the process proceeds to step S52 and the data rate is 9600b.
The reliability is determined to be 1 in ps, and the process ends.

If any of the comparison results in step S51 is not true, the process proceeds to step S53, and the threshold value set of the threshold value table shown in Table 4 stored in the threshold value storage unit 27 (FIG. 11) is set. Of these, the threshold value set of the threshold value set 2 which is the second highest reliability level of the threshold value set of 9600 bps with the highest occurrence probability among 9600 bps and 4800 bps with the correct CRC calculation result, and normalization of each of the four data rates The number of symbol errors and the average value of the path metric amount are compared.

If all the comparison results in step S53 are true, the process proceeds to step S54, the data rate is 9600 bps, the reliability is determined to be 2, and the process is terminated.

If any of the comparison results in step S53 is not true, the process proceeds to step S55, and the next 9600 bps of 9600 bps and 4800 bps of which the CRC calculation result is correct in the threshold value set of the threshold value table shown in Table 4. With a high probability of occurrence of 4800bp
The threshold value set 3 of s is compared with the normalized symbol error number and the path metric amount average value for each of the four data rates.

If all the comparison results in step S55 are true, the process proceeds to step S56, the data rate is 4800 bps, the reliability is determined to be 1, and the process ends.

If any of the comparison results in step S55 is not true, that is, in the threshold value set of the threshold value table shown in Table 4, the CRC calculation result is correct 9600b.
If none of the threshold set 1 and 2 or 3 of 4800 bps and the normalized symbol error number and the path metric amount average value of each of the four data rates are satisfied, the following step S57,
In S59, as in the case of FIG. 13 (or FIG. 14) described above, the threshold set and 4 are set in the order based on the probability of occurrence of the data rate that the traffic channel can take.
The normalized symbol error number and the path metric amount average value for each of the two data rates are compared.

Then, based on the comparison result, the data rate is 1200 bps and the reliability is 1 (step S5).
9), the data rate is 2400 bps, the reliability is 1 (step S60), or the determination of the data rate is impossible (step S61), and the process ends.

Further, in the rate determination processing of step S5 (FIG. 12), neither of the CRC calculation results of 9600 bps and 4800 bps was correct, and therefore, FIG.
Steps S71, S73, S75, S77, S7 shown in
At 9, the threshold value set is compared with the normalized symbol error number and the path metric amount average value of each of the four data rates in the order based on the probability of occurrence of the data rate that the traffic channel can take and its reliability.

Then, based on the comparison result, the data rate is 1200 bps and the reliability is 1 (step S7).
2), the data rate is 9600 bps and the reliability is 1 (step S74), the data rate is 9600 bps and the reliability is 2 (step S76), and the data rate is 2400 b.
If the reliability is 1 for ps (step S78), the data rate is 4800 bps and the reliability is 1 (step S80), or the data rate cannot be determined (step S81), the process ends.

As described above, the comparison / determination unit 41 (see FIG.
The determination result of the data rate processed and output in 1) is output to the decoded data storage unit 9, and the same processing as in the case of the first embodiment is performed thereafter.

As described above, in the second embodiment, the symbol error number is calculated excluding the symbol corresponding to the puncture symbol, and the puncture symbol, CRC code, and Since it is designed to multiply by the coefficient to correct the effect of the tail bit, it corresponds to the actual data,
It is possible to obtain a highly accurate normalized symbol error number, and as a result, the data rate can be determined more accurately.

Further, in response to the correctness of the CRC calculation result,
The number of normalized symbol errors for each of the four data rates,
Since the order of the threshold value set to be compared with the average value of the path metric amount is changed, when the CRC calculation result is actually correct, the data rate can be determined with a small number of comparisons, that is, quickly.

Further, even if the CRC calculation result is incorrect, the comparison process using the threshold value set of each of the four data rates is sequentially performed, so that the erroneous determination of the data rate can be prevented.

Further, since the comparison process is performed using the threshold value set in the descending order of the probability of occurrence of the data rate, the data rate can be determined more quickly.

The case where the present invention is applied to a mobile wireless device has been described above. However, the present invention can be applied to a decoding device for decoding a signal with a variable data rate, in addition to the mobile wireless device. it can.

In the second embodiment, the threshold value table of Table 4 shows only the threshold value set of 9600 bps and the reliability values of plural (1 and 2). A threshold set corresponding to the reliability can be used.

Further, in the first and second embodiments, the threshold table to be stored in the threshold storage unit 27 is determined when the sync channel is received. However, when the traffic channel is received, the threshold table is stored in the threshold storage unit 27. The threshold table to be set can be determined.

That is, immediately after shifting from the sync channel to the traffic channel, as shown in FIG. 17, a plurality of frames of a fixed data rate (for example, 1200 bps) are transmitted from the transmitting side (partner station side). When this is correctly received by the mobile wireless device (own station), a plurality of frames having a fixed data rate (for example, 9600 bps) are transmitted to the partner station side,
The communication path is established by correctly receiving this at the partner station side.

Therefore, while a plurality of frames having a fixed data rate such as 1200 bps are being transmitted from the partner station side, the threshold value table to be stored in the threshold value storage unit 27 is determined in the same manner as when receiving the sync channel described above. You can

In this case, since the threshold value table used for the comparison process is determined when the traffic channel actually used for the call is received, the data rate can be determined more accurately.

In the first or second embodiment, when the sync channel is received, the threshold table to be stored in the threshold storage unit 27 is determined, and then the comparison process is performed using the threshold table as it is. While receiving the traffic channel, eg, periodically, Eb / N
It is possible to supply 0 to the threshold setting unit 23 or 42 and change the threshold table used in the comparison process.

Further, in the second embodiment, the symbol error number is calculated excluding the symbol corresponding to the puncture symbol, but the normalized symbol number calculated from the symbol error number obtained by including this When the error is within the allowable range, the symbol error number can be calculated without removing the symbol corresponding to the puncture symbol. In this case, it is not necessary to provide the puncture symbol detection unit 31 (FIG. 10), and thus the device can be made compact.

[0163]

According to the decoding device of the first aspect,
The maximum likelihood decoding means performs maximum likelihood decoding of the received data based on the Viterbi algorithm. Then, the variable data rate of the second channel is determined based on the output of the maximum likelihood decoding means. Therefore, the data rate can be easily determined.

According to the decoding device of the second aspect, the number of symbol errors is calculated based on the output of the maximum likelihood decoding means, and one of the symbol error number and the output of the maximum likelihood decoding means is calculated. The data rate of the second channel is determined based on the path metric amount that is Therefore, it is possible to accurately determine the data rate.

According to the decoding device of the third aspect, the second channel includes the error detection code, and the error detection is performed based on the error detection code output by the maximum likelihood decoding means.
Then, the data rate of the second channel is determined based on the error detection result, the number of symbol errors, and the path metric amount. Therefore, the data rate can be determined more accurately.

According to the decoding device of the fourth aspect, the threshold value of each of the symbol error number and the path metric amount is set for each data rate that the second channel can take, based on the output of the maximum likelihood decoding means. Then, this threshold value is compared with the number of symbol errors and the path metric amount. Then, the data rate of the second channel is determined based on this comparison result. Therefore, the determination of the data rate is
It can be done easily.

According to the decoding device of the fifth aspect, the threshold value setting means sets the threshold value based on the output of the maximum likelihood decoding means corresponding to the first channel. Therefore, since there is a time margin from setting the threshold value to determining the data rate, it is possible to prevent a heavy load on the device.

According to the decoding device of the sixth aspect, the threshold value setting means sets the threshold value based on the output of the maximum likelihood decoding means corresponding to the second channel. Therefore, since the threshold value is set immediately before the data rate is determined,
It is possible to accurately determine the data rate.

According to the decoding device of the seventh aspect, the threshold value is set for each reliability. Then, the number of symbol errors and the path metric amount are sequentially compared with those having a high reliability and one having a higher data rate, and then the one having a lower reliability by one. The rates are compared in sequence from the highest to the lowest.
Therefore, when the number of symbol errors and the amount of path metric correspond to the threshold value of a high reliability and a high data rate, the determination result of the data rate can be obtained immediately.

According to the decoding device of the eighth aspect, the threshold value setting means sets the threshold value for each reliability. When the disconnection standard estimated when the data rate of the second channel is determined is worse than the disconnection standard estimated when the threshold setting means sets the threshold,
The reliability of the threshold applied to determine the data rate of the second channel is reduced and output. Therefore, the reliability of the determination result of the data rate is known, and when the reliability is low, the determination result can be prevented from being used, and thus, for example, malfunction of the device can be prevented.

According to the decoding device of the ninth aspect, the number of symbol errors and the path metric amount are compared based on the error detection result of the error detecting means for each possible data rate of the second channel. The order of the thresholds set in is determined. Therefore, the data rate can be determined quickly and more accurately.

According to the decoding device of the tenth aspect,
The number of symbol errors and the amount of path metric are compared in order from the threshold value corresponding to the data rate that the second channel is likely to take. Therefore, the data rate can be determined more quickly.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of a mobile wireless device to which a decoding device of the present invention is applied.

FIG. 2 is a diagram showing formats of a sync channel and a traffic channel.

FIG. 3 is a data rate determination unit 10 in the embodiment of FIG.
3 is a more detailed block diagram of FIG.

FIG. 4 shows a traffic channel with a data rate of 96.
It is a figure which shows the average value of the path metric amount vs. the average value of the normalized symbol error numbers when fixed at 00 bps.

FIG. 5 shows a traffic channel with a data rate of 48.
It is a figure which shows the average value of the path metric amount vs. the average value of the normalized symbol error numbers when fixed at 00 bps.

FIG. 6 shows a traffic channel with a data rate of 24.
It is a figure which shows the average value of the path metric amount vs. the average value of the normalized symbol error numbers when fixed at 00 bps.

FIG. 7: Data rate of 12 on traffic channel
It is a figure which shows the average value of the path metric amount vs. the average value of the normalized symbol error numbers when fixed at 00 bps.

FIG. 8 is a diagram showing bit positions of decoded data necessary for calculating a path metric amount average value per frame for four types of data rates of traffic channels.

9 is a flowchart illustrating an operation of a comparison unit 24 in FIG.

FIG. 10 is a block diagram showing the configuration of a second embodiment of a mobile wireless device to which the decoding device of the present invention is applied.

11 is a more detailed block diagram of the data rate determination unit 32 in the embodiment of FIG.

12 is a flowchart illustrating the operation of the comparison / determination unit 41 of FIG.

13 is a more detailed flowchart of step S2 in the flowchart of FIG.

FIG. 14 is a more detailed flowchart of step S3 of the flowchart of FIG.

15 is a more detailed flowchart of step S4 in the flowchart of FIG.

16 is a more detailed flowchart of step S5 of the flowchart of FIG.

FIG. 17 is a diagram illustrating a communication procedure in a traffic channel.

[Explanation of symbols]

 1 Viterbi Decoding Unit 2 Hard Decision Symbol Storage Unit 3 CRC Error Detection Unit 4 CRC Storage Unit 5 Convolutional Encoder 6 Symbol Error Number Calculation Unit 7 Symbol Error Number Storage Unit 8 Path Metric Amount Storage Unit 9 Decoded Data Storage Unit 10 Data Rate Determination Part 11 Channel selector 21 Normalization calculation part 22 Average value calculation part 23 Threshold value setting part 24 Comparison part 25 Judgment part 26 Eb / N0 estimation part 27 Threshold storage part 31 Puncture symbol detection part 32 Data rate judgment part 41 Comparison judgment part 42 Threshold value Setting unit 43 Normalization calculation unit

Claims (10)

[Claims] 1. A decoding device having maximum likelihood decoding means for performing maximum likelihood decoding of received data based on the Viterbi algorithm, wherein the received data has a fixed first channel and a variable second channel. And a rate determining means for determining the data rate of the second channel based on the output of the maximum likelihood decoding means. 2. An error number calculation means for calculating the number of symbol errors based on the output of the maximum likelihood decoding means, wherein one of the outputs of the maximum likelihood decoding means is a path metric amount, The decoding device according to claim 1, wherein the rate determination means determines the data rate of the second channel based on the number of symbol errors and the path metric amount. 3. The second channel includes an error detection code, further comprising error detection means for performing error detection based on the error detection code output from the maximum likelihood decoding means, and the rate determination means 3. The decoding device according to claim 2, wherein the data rate of the second channel is determined based on an error detection result of the error detecting means in addition to the number of symbol errors and the path metric amount. 4. Further comprising threshold setting means for setting a threshold for each of the number of symbol errors or the path metric amount for each data rate that the second channel can take on the basis of the output of the maximum likelihood decoding means, 3. The rate determining means compares the threshold with the number of symbol errors and the path metric amount, and determines the data rate of the second channel based on the comparison result. Decoding device according to item 3. 5. The decoding apparatus according to claim 4, wherein the threshold setting unit sets the threshold based on an output of the maximum likelihood decoding unit corresponding to the first channel. 6. The decoding apparatus according to claim 4, wherein the threshold setting unit sets the threshold based on an output of the maximum likelihood decoding unit corresponding to the second channel. 7. The threshold setting unit sets the threshold for each reliability, and the rate determining unit sets the symbol error number and the path metric amount to a threshold with high reliability and a data rate. Is sequentially compared with one from a higher to a lower one, and then from a higher to a lower data rate, which has a lower reliability by one. 7. The decoding device according to any one of 6 to 6. 8. The threshold value setting means sets the threshold value for each reliability thereof, and the rate determining means determines a line disconnection standard estimated when the data rate of the second channel is determined, When the threshold setting unit is worse than the line disconnection standard estimated when the threshold is set, the reliability of the threshold applied to determine the data rate of the second channel is lowered and output. The decoding device according to any one of claims 4 to 6. 9. The rate determining means compares the number of symbol errors and the path metric amount based on the error detection result of the error detecting means, and sets for each possible data rate of the second channel. 7. The decoding apparatus according to claim 4, wherein the order of the thresholds determined is determined. 10. The rate determining means compares the number of symbol errors and the path metric amount in order from a threshold value corresponding to a data rate that is likely to be taken by the second channel. 7. The decoding device according to any one of 4 to 6.