JPS612439A - Digital signal transmission system - Google Patents

Abstract

PURPOSE:To obtain easily an error correction effect almost equivalent to that of a low encoding ratio code of which encoding ratio is 1/m that of an original code through a simple encoder/decoder by decoding equivalently a code having k/nm encoding ratio to a code having k/n encoding ratio by m-times of multiplex-transmission at transmission side and averaging operation at reception side. CONSTITUTION:An input data string 1 having R data speed is converted into an encoded data string 3 by an original code encoder 2 and the data string 3 is sent to an m-times multiplexer 4 for an encoded bit. At that time, the data speed is n/k times which is the reciprocal of k/n encoding ratio. A transmission data string 5 having the data speed of m-times that of the encoded data string 3 is formed by multiplexing each encoded bit included in the data string 3 repeatedly m times through the m-times multiplexer 4 and a transmission signal 7 modulated by a modulator 6 is sent to a transmission line. On the receiving side, a base band signal sample value 10 obtained by demodulating a receiving signal 8 by a demodulator 9 is averaged by an m-bit averaging circuit 11 and a decoder input data string 12 to be inputted to an original code decoder 13 is formed on the basis of said average value.

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention relates to a digital communication system that performs error correction using error correction codes, and is particularly applicable to systems that require improvement in signal transmission quality. This invention relates to an effective digital signal transmission method.

(Prior Art) Introducing error correction technology into digital communication systems is an effective means to improve and guarantee the quality of transmitted signals. Conventionally, when the receiving side detects an error in the transmitted signal, it requests the transmitting side to retransmit the signal (ARQ).
(tomatic Repeat Request) method has been widely adopted. However, in recent years, FEC (Forward E
RRCorrection) methods are being actively studied, and their practical application is also progressing in various fields.

Adding redundancy to a transmission signal is called error correction coding, and generally when error correction coding is performed, the amount of data to be transmitted increases due to the redundancy. Therefore, in order to keep the amount of information transmitted per unit time constant, it is necessary to increase the overall data transmission rate, and conversely, if the data transmission rate is limited, the information transmission rate must be lowered. .

On the other hand, the reciprocal of the rate of increase in redundancy due to error correction coding is called the coding rate, and generally speaking, the lower the coding rate, the greater the error correction capability.

From the above, the coding rate of the error correction code to be applied to the communication system needs to be determined by comprehensively considering the required error rate improvement effect and the constraints on the signal transmission band. However, error correction codes with a coding rate of 1/2 (one bit of the original data is encoded into 2 bits with added redundancy) or a relatively high coding rate are usually used. Often. This means that even if an error correction code with a coding rate lower than 1/2 is applied, the amount of noise will increase as the signal band increases due to coding, and the error rate will be low enough to compensate for the expansion of the transmission band. This is because no improvement effect can be expected.

On the other hand, in data communications, depending on the importance of the information to be transmitted, highly reliable information may be required using existing channels. Conventionally, in such cases, error propensity has been significantly improved by allowing the information transmission speed to decrease and by applying a code with a low coding rate.

(Problem to be solved by the invention) However, in systems where high quality transmission is required only for specific information, multiple error correction devices with different coding rates are prepared, and It is necessary to select the equipment to be used depending on the required quality, which complicates the entire communication system and is also economically problematic. Furthermore, low coding rate error correction codes themselves are not well known, and there are restrictions on the selection of error correction codes, and there are also limits to the effect of improving the error rate.

In addition, no technology has been established to harden encoders and decoders for such error correction codes. The present invention was developed in view of the above-mentioned drawbacks of the prior art, and uses an error correction code with a relatively high coding rate to achieve the same error improvement effect as when applying an error correction code with a low coding rate. The purpose is to provide a simple digital signal transmission system that can

(Means for solving the problem) A feature of the present invention is that on the transmitting side, the coding rate is
When there are n known codes, by multiplexing the encoded bits m times, a code with a low encoding rate equivalent to /nm is generated, and on the receiving side, the code is multiplexed m times. A decoder input data sequence is generated based on the average value of the demodulated baseband signal of the sample values of the demodulated signal of each bit, and is input as is to a decoder with a code of /n as the coding rate.

(Operation) The present invention can equivalently decode a code with a coding rate of /nm for a code with a coding rate of /n by multiplexing m times on the transmitting side and averaging operation on the receiving side. can.

(Example) FIG. 1 is a diagram showing the principle of a signal transmission system based on the present invention. An input data sequence l having a data rate of R is converted into a coded data sequence 3 by an encoder 2 for the original code, and is sent to an m-concave weighting circuit 4 for coded bits. The data rate at this time is twice the coding rate as the reciprocal of /n. The m-concave line weighting circuit 4 generates a transmission data sequence 5 having a data rate m times that of the encoded data sequence 3 by repeating each encoded bit included in the encoded data sequence 3 m times. The data sequence 5 is modulated by a modulator 6 and sent out as a transmission signal 7 to a transmission path. On the other hand, on the receiving side, the baseband signal sample value 10 obtained by demodulating the received signal 8 with the demodulator 9 is transmitted to the m-bit averaging circuit 1 corresponding to the m-concave weighting circuit 4 on the transmitting side.
Sent to 1. The m-bit averaging circuit 11 averages the demodulated baseband signal sample values lO corresponding to the m bits repeatedly sent out on the transmitting side, and inputs the average value to the original code demodulator 13 based on the average value. A decoder input data sequence 12 based on hard decisions or soft decisions is generated.

The decoder input data series 12 is decoded by the original frame decoder 13, and decoded data 14 is extracted.

FIG. 2 shows a specific example of the configuration of the transmission system. The symbols in the figure are the same as in FIG. 1. Repeatedly sending the same data m times on the transmitting side is a type of speed conversion. Therefore, in the configuration example shown in Figure 2, the clock speed is basically (n/11
Two types of clocks CL are R) and (nIl/ni・R).
, , Prepare Cl3. In this way, the m-column weighting circuit 4 can be constituted by a simple latch circuit. That is, the coded sequence 3 from the original code encoder 2 is processed by a clock CL having a speed equal to the data rate (n/R) of this sequence,
Each data is held for the period of the clock OL. Since this held data is always output to the modulator 6, if the modulator 6 is operated with a clock CL2 having a speed of (nm/km R), each data is
This means that it has been sent repeatedly.

On the other hand, FIG. 3 shows a specific configuration example of the receiving system. In the figure, 15 is an A/D converter, 17 is an adder, and 18 is an R
OM, 18.18 indicates a signal, and the rest is the same as in FIG.

The baseband sample value (analog value) 10 of the received signal output from the demodulator 9 is quantized by the A/D converter 15 and becomes a signal 1B, which is input to the adder 17. The adder 17 adds the quantized value 18 of m-bit baseband signal samples, and the value 18 is stored in the ROM 19.
is output to. Functionally, the ROM 19 has a 1/++ division function and a demodulation data 12 creation function,
The data written in this ROM19 is stored in the adder 1
7 as an address, hard decision data or soft decision data predetermined by the characteristics of the decoder 13 corresponding to the value is written. The decoder 13 is RO
A decoding operation is performed based on the data 12 sequentially read from M19.

Note that when adding m bits, the adder 17 needs to operate in synchronization with the m bits multiplexed on the transmitting side. The signal for this is the Re5et signal, and this signal is the synchronization signal sent from the transmitting side or the decoder 13
It is generated based on the synchronization information etc. extracted in .

Furthermore, the demodulator 9, A7D converter 15, and adder 17 basically operate with a clock cL2 at a speed of (nm/R), and the ROM 19 and decoder 13 operate with a clock cL2 at a speed of (n/kIIR). CL, but these clocks CL and Cl3 must be synchronized with the above-mentioned Re5et signal.

In addition, in FIG. 2, if the original code encoder 2 has a latch function for the output signal, there is no need to provide a latch circuit as the m-multiplexing circuit 4, and in FIG. If the A/D conversion function is provided, the A/D converter 15 is not necessary.

The above is an embodiment of the signal transmission system based on the present invention. If the coding rate of the original code is /n and the data rate of input data sequence 1 is R, the data rate of encoded data sequence 3 is n.
/kR, the data rate of transmission data series 5 is (nm)/k
It becomes R. That is, the transmission data series 5 is the input data series l
This is equivalent to encoding the input data sequence l at a coding rate of /(nm).

Assuming that the signal transmission method based on the present invention is applied to the communication channel where white Gaussian noise is added, the energy per 1 bit of the transmission signal 7 is Es, and the variance of Gaussian noise is No/2.
(No: one-sided noise power density). When the m-bit averaging operation of the demodulated baseband signal 1° is performed using the hydraulic method, the energy Es per 1 bit of the signal does not change, but
Since the dispersion of noise is multiplied by 1/m, Es/N of the transmission path.

is equivalently mE in the decoder input data series 12.
The result will be improved to s/No. That is, the decoded bit error rate vs. Es/No characteristic of the original code is +01og+a m
(dB), and almost the same error immediacy as when a low coding rate code of /(nm) is directly decoded can be obtained.

Here, an example of the effect of improving the bit error rate when the method based on the present invention is applied will be shown, taking as an example a communication system in which error correction is performed by soft-decision and Itabi decoding using convolutional codes. As the original code, the constraint length = near and the coding rate is 1/
Assume that we use a convolutional code of 2. In convolutional codes used for maximum likelihood decoding such as Tibi decoding, special algebraic properties are not required for the code, and it is said that the larger the minimum distance d of the code, the greater the error rate improvement effect during decoding. It has characteristics. Therefore, the data sequence 5 obtained by multiplexing the original code differentiation bits m times based on this method is a coded sequence having a distance m times that of the original code.

Since the minimum distance lad of a code with a coding rate of 1/2 when K=7 is 10, the transmission data sequence 5 after multiplexing (coding rate of 1/2 (encoded sequence) is 10 m. Table 1 shows that m=
The code generation polynomial of the 1/4 code obtained as
nvolutional codes with
maximal freedom fo
r rates 1/2.1/3 and 1/4,
IEEE Trans, on Inf, Theo
ry, IT-19, pp, 371. May197
3), and the minimum distance d between the codes is 20 for all codes.

FIG. 4 shows a comparison of the decoded bit error rate (BER) versus Es/No characteristics during 8-value soft-decision Viterbi decoding of these two types of codes with a coding rate of 1/4. however,
The characteristics of the code based on the method based on the present invention are compared with the characteristics of the original code with a coding rate of 1/2 (K = 7) shown in the same figure.
It is calculated based on the assumption that s/N o is 10 log and o 2 =3 dB less. The difference in the characteristics of the inner codes shown in FIG. 4 is very small, and the low coding rate code obtained by the method based on the present invention can provide practically sufficient error rate improvement characteristics.

Ideally, the m-times averaging operation of the demodulated baseband signal sample value lO in Fig. 1 should be performed based on the analog value, but after A/D (analog/digital) conversion of the sample value 10, Even if an averaging operation is performed in digital processing, if the number of quantization bits is sufficiently large, the resulting deterioration in error patternability can be ignored. In particular, when the decoder for the original code handles soft-decision data, the demodulator usually incorporates a linear A/D converter with a much larger number of bits (about 8 bits) than the number of soft-decision bits. Therefore, by performing an averaging operation based on this output signal, the m-bit averaging circuit 11 can be easily realized.

Table 12 types of code parameters (Coding rate 114. Constraint length = 7) (Other examples) Next, other embodiments of the present invention will be described.

In the principles and embodiments described in FIGS. 1, 2, and 3, m-bit multiplexing is performed on the time axis, but it is not necessarily necessary to perform it on the time axis.

For example, if the modulation method is 4-phase PSK, P channel, Q
The data of two channels of channels are multiplexed. Therefore, if two bits are multiplexed using the m-groove multiplexing circuit 4 in Fig. 1 and one of the multiplexed two bits is assigned to the P channel and the other to the Q channel, it becomes equivalent to two-phase PSK transmission. . If the receiving side demodulates with 2-phase PSK instead of 4-phase PSK, it is possible to obtain the input data sequence 12 to the decoder 13 equivalent to the present invention without performing 2-bit averaging operation.

It goes without saying that multiplexing operations can also be performed spatially or on the frequency axis using the same concept as above.

Next, an example will be described in which the present invention is applied to a system that performs route diversity.

In a system that performs signal reception using route diversity, the transmitting side does not have to repeatedly send coded bits, but the receiving side has an independent demodulator for each receiving route and averages the output base punt signals. , when there is no significant difference between the transmission conditions of both routes, the same error rate improvement effect as the method based on the present invention can be expected.

FIG. 5 shows an example of the configuration of a receiving system when the present invention is applied to route diversity. In the figure, 201
~20m is a reception route, 91~9■ is a demodulator provided for each reception route, 21,~21m is a demodulator 91
This is a latch circuit that holds the outputs of the 9th layer for a certain period of time, and the rest is the same as in FIG. 3. It is assumed that the demodulators 91 to 9 include an A/D converter that converts demodulated baseband signal sample values into digital signals.

each roux) 20. A data sequence with a coding rate of /n is sent from the ~2 axes at a data rate of n/kIIR. If the adder 17 calculates the sum of the baseband signal sample values 101 to 10m obtained by demodulating this for each route, the output signal 18 of the adder 17 will have the same effect as that in FIG. 3.

(Application example) Next, an application example of the present invention will be described.

Based on the present invention, a low coding rate method for error correction codes and a high coding rate method using punctured coding for convolutional codes (Yasuda, Hirata, Ogawa: “Easy high coding rate for Vitabi decoding”) Rate convolutional codes and their characteristics, Transactions of the Institute of Electronics and Communication Engineers B, J84-B, ?, pp.

573-580. July 1982) is combined, 1
A coding rate variable error correction device that can freely change the coding rate can be easily configured using the same device. Figure 6
is a basic block diagram of such a device. As shown in the figure, data encoded by the original code encoder 2 with a coding rate of /n has some of the encoded bits erased by the encoded bit erasure circuit 22 (punctured encoding). ),
Based on the present invention, the encoded sequence 23 with a coding rate of '/n'(k'/n'>k/n) is sent to the <m-concave line weighting circuit 4. The coding rate of the equivalent code written as a result is′
/an'. The reverse operation is performed on the receiving side, and the dummy data insertion circuit 28 for punctured code decoding is used on the transmitting side! Dummy data is inserted into the erased bit positions to create a data series 12 corresponding to the coding rate series of the original code.
is reproduced and decoded by the decoder 13 with a coding rate of /n for the original frame.

With such a device configuration, by simply adding a simple peripheral circuit to the encoder/decoder 2 for original frames with a coding rate of /n, the number and pattern of deletion/insertion bits for punctured encoding can be changed Based on the present invention, by controlling the number of times of multiplexing (m), the coding rate can be freely changed (it can be made higher or lower than the coding rate of the original code). A variable error correction device can be realized.

(Effects of the Invention) By using the method of the present invention, by using a general-purpose or relatively simple code/decoder as is, errors can be reduced to the same level as a low coding rate code with a coding rate of 1/m of the original code. Since the correction effect can be easily obtained, it is particularly effective when performing highly reliable data transmission using an existing communication channel. In addition, in systems that originally require wideband transmission, such as spread spectrum systems for the purpose of protection from jamming waves and interference waves, the coding rate of the code is reduced using the tree-nut method while keeping the information transmission speed constant. It is effective to expand the signal transmission band by doing so.

[Brief explanation of the drawing]

Figure 1 is a diagram showing the principle of a signal transmission system based on the present invention, Figure 2 is a block diagram of a transmitting system according to the present invention, Figure 3 is a block diagram of a receiving system according to the present invention, and Figure 4 is a diagram showing decoding bits. FIG. 5 is a block diagram of another embodiment of the present invention. FIG. 6 is a block diagram of still another embodiment of the present invention. 2: Encoder for original code, 4: M-concave line weighting circuit. 6: Modulator, 9: Demodulator. 11: m-bit averaging circuit, 13: original code decoder.

Claims (1)

[Claims] The transmitter side has at least an encoder and a modulator for error correction codes, and the receiver side has at least a demodulator and a decoder for error correction codes, and the error correction code is used between the transmitter and the receiver. In a digital signal transmission method for transmitting signals, the transmitting side repeatedly transmits the output bits of the decoder multiple times, and the receiving side averages the demodulated baseband signals of the demodulator corresponding to the repeatedly transmitted bits. and generates an input data sequence to the decoder based on the value obtained by averaging, and performs encoding that is equivalently lower than a predetermined encoding rate for the encoder and decoder. A digital signal transmission method characterized by transmitting digital signals at different rates.