JP5195989B2 - Sending method - Google Patents

Description

  The present invention relates to a transmission method, and in particular, inserts dummy bits into information bits and encodes to generate a systematic code, deletes the dummy bits from the systematic code, transmits the dummy bits, and receives a dummy bit with the maximum likelihood on the receiving side. It is related with the transmission method in the system which inserts and decodes.

As shown in FIG. 35, when an N-bit code I ₂ is generated by encoding information bits I ₁ composed of K bits, a code in which K bits of the codes are composed of original information is generated. It is called a systematic code, and the remaining M (= N−K) bits are called parity. An example of the system code is a turbo code.

Consider the information alphabet as a general form of bits. One alphabet has q kinds of symbols {a ₀ , a ₁ , a ₂ , .. a _q-1 } as values, and the bit is a special case of q = 2, and a ₀ = 0, , a ₁ = 1.
On the transmission side, K information alphabets u = (u ₀ , u ₁ ,..., U _K−1 ) and K × N generator matrix G = (g ij); i = 0,. . . , K−1; j = 0,. . . , N-1
Using the following formula: x = uG
If N code alphabets x = (x ₀ , x ₁ ,..., X _N-1 ) are generated, the code alphabet becomes a block code, and the information alphabet u is block-coded.
On the receiving side, the information alphabet u is estimated from the received data for the code vector x. For this purpose, the following parity check relation xH ^T = 0 for x
Is used. here,
H = (hij); i = 0,. . . , M−1; j = 0,. . . , N−1
In the parity check matrix, H ^T denotes the transpose of H (interchanging rows and columns). From the above two equations, H and G satisfy the following relationship.
GH ^T = 0
From this, when one of H and G is given, the encoding rule is uniquely determined.

FIG. 36 is a block diagram of a communication system that performs block encoding at the transmitter and decodes at the receiver. The transmitter 1 encodes information u consisting of K bits to generate an N-bit block code x, and a coding unit 1a. A modulation unit 1b that modulates and transmits the block code is provided. The receiver 2 includes a demodulator 2a that demodulates a signal received via the transmission path 3, and a decoder 2b that decodes the originally transmitted K-bit information from the N-bit received information.
The encoding unit 1 combines a parity generator 1c for generating M (= N−K) parity bits p, K-bit information u and M-bit parity bits p, and N (= K + M) block codes x. The P / S conversion unit 1d is provided. For example, a turbo code can be used as the code of the encoding unit 1a. The decoding unit 2b includes a decoder 2c that performs error detection and correction processing on the reception likelihood data y, decodes the original transmitted K-bit information, and outputs estimated information. The block code x transmitted from the transmitter 1 is not input as it is transmitted to the decoder 2c under the influence of the transmission path 3, but is input to the decoder 2c as likelihood data. Likelihood data consists of the reliability of whether the sign bit is 0 or 1, and the sign (0 if +1, 1 if −1). The decoder 2c performs a prescribed decoding process based on the likelihood data for each code bit, and estimates the information bit u. In the case of a turbo code, the decoder 2c performs maximum a posteriori probability decoding (MAP decoding).

FIG. 37 is a block diagram of the turbo encoder 1a, and FIG. 38 is a block diagram of the turbo decoder 2b. A turbo code is a systematic code composed of several element encoders and an interleaver. By adopting MAP decoding, errors in decoding results can be reduced each time the number of decoding iterations is increased.
FIG. 37 shows an example in which two element encoders are arranged in parallel across one interleaver, and u = [u0, u1, u2, u3, .., u _K-1 ] is Information data of length K to be transmitted, xa, xb and xc are encoded data obtained by encoding the information data u by the turbo encoder 1a, and ya, yb and yc are the encoded data xa, xb and xc through the communication path 3. A received signal propagated and influenced by noise and fading, u ′, is a decoding result obtained by decoding the received data ya, yb, yc in the turbo decoding unit 2b. In the turbo encoder 1a, the encoded data xa is the information data u itself, the encoded data xb is the data obtained by convolutionally encoding the information data u by the first element encoder ENC1, and the encoded data xc is the information data u. The data is interleaved (π) and convolutionally encoded by the second element encoder ENC2. That is, the turbo code is a systematic code synthesized by using two convolutions, where xa is an information bit and xb and xc are parity bits. The P / S converter 1d converts the encoded data xa, xb, xc into serial data and outputs it.
In the turbo decoding unit 2b of FIG. 38, the first element decoder DEC1 performs decoding using ya and yb among the received signals ya, yb and yc. The first element decoder DEC1 is a soft decision output element decoder and outputs the likelihood of the decoding result. Next, the second element decoder DEC2 performs similar decoding using the likelihood and yc output from the first element decoder DEC1. The second element decoder DEC2 is also an element decoder with a soft decision output, and outputs the likelihood of the decoding result. In this case, since yc is a received signal corresponding to xc obtained by encoding the interleaved original data u, the likelihood output from the first element decoder DEC1 is before input to the second element decoder DEC2. Interleave (π). The likelihood output from the second element decoder DEC2 is deinterleaved (π- ¹ ) and then fed back as an input to the first element decoder DEC1. Note that the result of the hard decision of “0” and “1” for the deinterleave result of the second element decoder DEC2 is the turbo decoding result (decoded data) u ′. Thereafter, the error rate of the decoding result u ′ is reduced by repeating the above decoding operation a predetermined number of times. MAP element decoders can be used as the first and second element decoders DEC1 and DEC2 in the turbo decoding unit.

As a specific form of the communication system of FIG. 36, a 3GPP W-CDMA mobile communication system can be considered. FIG. 39 is a block diagram of a 3GPP W-CDMA mobile communication system, in which the radio base station is the transmitter of FIG. 36 and the mobile station is the receiver. 39, the mobile communication system includes a core network 11, radio base station controllers (RNCs) 12, 13, demultiplexers 14, 15 and radio base stations (Node B) 16 _{1 to} 16 ₅ , mobile It consists of a station (UE: User equipment) 17.
The core network 11 is a network for performing routing in the mobile communication system. For example, the core network can be configured by an ATM switching network, a packet switching network, a router network, or the like. The core network 11 is also connected to other public networks (PSTN) or the like, and the mobile station 7 can also communicate with a fixed telephone or the like.
The radio base station controller (RNC) 12, 13 are positioned as higher-level devices of the wireless base station 16 ₁ to 16 _5, these control of the radio base station 16 ₁ to 16 ₅ (management of wireless resources used, etc.) Has the ability to do. The demultiplexers 14 and 15 are provided between the RNC and the radio base station. The demultiplexers 14 and 15 demultiplex a signal addressed to each radio base station received from the RNCs 12 and 13 and output the signal to each radio base station. Performs control to multiplex the signals from the stations and deliver them to each RNC side.
The radio base stations 16 _{1 to} 16 ₃ perform radio communication with the mobile station 17 while the radio resources are managed by the RNC 12 and the radio base stations 16 ₄ and 16 ₅ are managed by the RNC 13. The mobile station 17 is located in the radio area of the radio base station 16, thereby establishing a radio line with the radio base station 16 and communicating with other communication devices via the core network 11. Do.

The above is a description of a general mobile communication system. In order to enable high-speed downlink data transmission (packet transmission), an HSDPA (High Speed Downlink Packet Access) system is adopted.
HSDPA adopts an adaptive coding modulation system. For example, the number of bits of the transport block TrBL, the number of multiplexed codes, and the modulation system (QPSK modulation system, 16-value QAM system) are wireless between the radio base station and the mobile station. It is characterized by adaptive switching according to the environment.
In addition, HSDPA employs an H-ARQ (Hybrid Automatic Repeat reQuest) method. In H-ARQ, when a mobile station detects an error in received data from a radio base station, it makes a retransmission request (transmission of a NACK signal) to the radio base station. The radio base station that has received this retransmission request retransmits the data, so that the mobile station performs error correction decoding using both the already received data and the retransmitted received data. As described above, in H-ARQ, even if there is an error, the already received data is effectively used, so that the gain of error correction decoding is increased, and as a result, the number of retransmissions can be reduced. When an ACK signal is received from a mobile station, data transmission is successful, so retransmission is unnecessary, and the next data transmission is performed.
As shown in Fig. 4, the main radio channels used for HSDPA are (1) HS-SCCH (High Speed-Shared Control Channel), (2) HS-PDSCH (High Speed-Physical Downlink Shared Channel), (3) There is HS-DPCCH (High Speed-Dedicated Physical Control Channel).
HS-SCCH and HS-PDSCH are common channels (shared channels) in the downlink direction, that is, in the direction from the radio base station to the mobile station, and HS-PDSCH is a common channel (shared channel) that transmits packets in the downlink direction. HS-SCCH is a control channel that transmits various parameters related to data transmitted by HS-PDSCH. In other words, it is a channel that notifies that data transmission is performed via HS-PDSCH, and the allocation number of destination mobile station information, transmission bit rate information, modulation scheme information, spreading code (spreading code) as various parameters ( The number of codes) and the rate matching pattern for the transmission data.
HS-DPCCH is an individual control channel (dedicated control channel) in the uplink direction, that is, from the mobile station to the radio base station, and the reception result (ACK signal, NACK signal) of data received via HS-PDSCH Is used when transmitting from a mobile station to a radio base station. The HS-DPCCH is also used to transmit a CQI (Channel Quality Indicator) based on the reception quality of the signal received from the radio base station to the radio base station. Based on the received CQI, the radio base station determines whether the downlink radio environment is good or not. If it is good, the radio base station switches to a modulation method capable of transmitting data at a higher speed. Is switched to a modulation scheme for transmitting, thereby performing adaptive modulation. Actually, the base station holds a CQI table that defines formats with different transmission rates according to CQI = 1 to 30, and the parameters (transmission rate, modulation scheme, number of multiplexed codes, etc.) according to CQI are set in the CQI. It is obtained from the table and notified to the mobile station by HS-SCCH, and data is transmitted to the mobile station by HS-PDSCH based on the parameter.
In the above 3GPP W-CDMA mobile communication system, the transmitter 1 in FIG. 36 is a radio base station, and the receiver 2 is a mobile station (mobile terminal).

FIG. 41 shows a data transmission processing block of the 3GPP W-CDMA radio base station, and FIG. 42 shows a data format for explaining the transmission processing (see Non-Patent Document 1). An example will be described in which the number of code blocks is 2, 1stRM and 2nsRM in the physical layer H-ARQ function unit are both puncturing, and the number of physical channel codes is 2.
(1) Information bits are passed from the upper layer as a transport block TB.
(2) The CRC adding unit 21 performs encoding for error detection by CRC (Cyclic Redundancy Check) for each transport block TB. That is, based on the transport block TB, a CRC parity bit having a specified number of bits is generated and added after the transport block TB itself. ... Data set D1
(3) Next, the bit scrambling unit 22 performs bit scrambling on the data set D1. Bit scrambling is a bit addition of a pseudo-random bit pattern B = (b0, ..., b (K-1)) generated by a specified generation method of the same size K as the data set D1 and the data set D1 (Hereafter, arithmetic operations between bits all mean mod2 operations on [0,1]). ... Data set D2
(4) Code Block Segmentation Unit 23 Code block segmentation is performed on the data set D2. That is, when the size K of the data set D2 exceeds the specified size Z, the data set D2 is divided into a plurality of code blocks all having the same data size. If the data is not divisible by the division number C, adjust by adding a filler bit. The value of the filler bit is set to 0 and added to the beginning of the original data. In the turbo code, Z ≦ 5114 because 40 ≦ K ≦ 5114. ... Data set D3
(5) The channel coding unit (encoding unit) 24 performs encoding on each code block of the data set D3. The code is a turbo code with a specified coding rate R = 1/3. ... Data set D4
(6) The physical layer HARQ functional unit 25 performs H-ARQ processing (H-ARQ Functionality) on the data set D5. The bit dividing unit 25a of the physical layer HARQ function unit 25 divides each code block output from the encoding unit 24 into systematic bits, parity bits 1, and parity bits 2, and serially connects the same ones. ... Data set D5
(7) The first rate matching unit 25b of the physical layer HARQ function unit 25 checks whether the total bit length of the data set D5 is larger than the specified buffer size NIR, and if so, sets the data set D5 so that the size becomes NIR. Puncture and do nothing when below NIR. Puncturing is performed on parity 1 and parity 2, not on systematic bits. ... Data set D61
Next, the second rate matching unit 25c of the physical layer HARQ function unit 25 performs rate matching (repetition or puncturing) on the data set D61 according to the designated H-ARQ transmission parameter. As H-ARQ transmission parameters,
・ Modulation method (QPSK or 16QAM)
・ All bit size Ndata of physical channel HS-PDSCH
・ H-ARQ transmission pattern RV,
and so on.
The total bit size Ndata is
Ndata = number of codes × physical channel size, and the physical channel size is 960 for QPSK and 1920 for 16QAM. When the size of the data set D61 is smaller than Ndata, the second rate matching unit 25c performs repetition so that the size of Ndata is obtained. When larger than Ndata, puncturing is performed. ... Data set D62
The repetition is a process of selecting a designated number from the code bits, creating a copy of the code bit, and adding the copy. On the receiving side, diversity combining is performed so that the data of the same bit improves SN. Puncturing is a process of selecting a specified number of bits from code bits and deleting the bits. On the receiving side, a fixed maximum likelihood value is added as bit data at the deletion position.
Among the above parameters, the modulation scheme, the number of codes, RV, etc. are reported to the receiver (terminal) through another common channel HS-SCCH.
(8) The bit combination unit 25d of the physical layer HARQ function unit 25 performs bit combination on the data set D62 and outputs a combination result. Next, the bit combination unit 25d replaces the data order in order to map the systematic bits and the parity bits into one modulation signal symbol.
This replacement process is a kind of interleaving. That is, the number of bits Ncol mapped to one modulation signal symbol is the number of columns, the number of rows is Nrow = Ndata / Ncol, and the bits of the data number Ndata are arranged in a matrix. Ncol = 2 for QPSK and Ncol = 4 for 16QAM. In the replacement process, the systematic bit arrangement area and the parity bit arrangement area are divided so that the systematic bit is located in the upper row. For example, for 16QAM, the systematic bits are preferentially mapped to the first 2 bits of the 4 bits by this arrangement process. This is because 16QAM mapping is determined so that the reliability of the likelihood of the first two bits is increased. A set of bits in each column of the array is one modulation signal symbol. ... Data set D7
(9) The physical channel separation unit 26 divides the data set D7 into physical channels (Physical
Channel Segmentation). The number of divisions is the number of codes. Data set D7 is serially divided from the first bit by the number of divisions. ... Data set D8
(10) The HS-DSCH interleaving unit 27 performs H-ARQ interleaving processing (H-ARQ Interleaving) on the data set D8. That is, the interleaving unit 27 performs interleaving on each physical channel with a prescribed interleaving pattern. ... Data set D9
(11) The constellation rearrangement unit 28 performs symbol rearrangement (Constellation Re-arrangement) on the data set D9 whose modulation scheme is 16QAM. However, nothing is done when the modulation method is QPSK. Symbol rearrangement performs bit substitution and inversion in units of 4 bits for each symbol according to specified parameters. ... Data set D10
(12) The physical channel mapping unit 29 maps the data set D10 to a physical channel (Physical Channel Mapping) and transfers the physical channel data of the data set D10 as it is to the modulation unit.

As an encoding / decoding method for systematic codes, in order to improve the error rate characteristics of decoding results, encoding is performed by inserting dummy bits into information bits on the transmission side, and dummy bits are deleted from the codes obtained by encoding. A technique for transmitting a systematic code is proposed (see Patent Document 1 and Patent Document 2). FIG. 43 is an explanatory diagram of the encoding / decoding method proposed in Patent Document 1. In FIG.
K0 dummy bits 200 having a predetermined pattern are inserted into K information bits 100 to form K1 (= K + K0) pieces of first information. The dummy bits repeat all 10 patterns or 1 and 0 alternately. . . Not only 10 patterns but a predetermined pattern can be used. The dummy bits 200 can be inserted into the information bits before or after the information bits 100 or evenly. In the figure, a dummy bit 200 is inserted after the information bit 100.
Next, M parity bits 300 created using the K1 information bits are added to the K1 information to generate N1 (= K1 + M) bits of information 400 (tissue coding). Thereafter, K0 dummy bits 200 are deleted from the information to generate an N (= K + M) -bit organizational code 500, the systematic code 500 is transmitted from the transmitter to the receiver, and decoded at the receiver. . Note that the coding rate R = K / (K + M).
The decoding unit of the receiver inserts the dummy bit 200 deleted on the transmitting side into the demodulated system code 500 with the maximum likelihood (reliability ∞), and then turbo-decodes and outputs the information bit 100.

FIG. 44 is a block diagram of a communication system for realizing the encoding / decoding method of FIG. 43, and the same components as those in FIG. The encoder 1a of the transmitter 1 applies forward error correction (FEC) to the information bit u to perform transmission with high reliability, and the modulator 1b modulates the resulting code bit x. And transmit to the receiver 2 through the wireless transmission path 3. The demodulator 2a of the receiver 2 demodulates the received data, and the likelihood data y consisting of the reliability of the code bit 0 or 1 and the hard decision code (+ 1 → 0, −1 → 1) is sent to the decoder 2b. input. The decoding unit 2b performs a prescribed decoding process based on the likelihood data for each code bit, and estimates the information bit u.
In the encoding unit 1a of the transmitter 1, the dummy bit inserting unit 1e inserts K0 randomly selected 0 and 1 bits into K information bits u as dummy bits at randomly selected positions, and K1 = K + K0 information bits (u, a) = (u 0, .., u K-1, a 0, .., a K0-1)
Is output. Encoder 1f outputs performs turbo coding N1 (= K + K0 + M ) information bits _{x 1 (u, a, p} ) and using the information bits of the K1 bits dummy bit is inserted. Where p is M parity bits,
p = (p ₀ ,..., p _M-1 )
It is. The dummy bit deletion portion 1g is, N1 pieces of data output from the encoder 1f bit _{x 1 (u, a, p} ) delete the K0 pieces of dummy bit a N information bits x = (u, p) = (X ₀ , x ₁ , ..., x _N-1 )
Is generated. Modulator 12 modulates this information bit x and transmits it.
The demodulator 2a of the receiver 20 receives and demodulates the noise-added data through the transmission path 3, and the likelihood data y = (y ₀ , y ₁ ,..., Y _{N-1 of} each code bit. )
Is input to the decoding unit 2b. The dummy bit likelihood inserting unit 2d of the decoding unit 2b inserts likelihood data a with the maximum likelihood (reliability ∞) into the dummy bit position inserted by the transmitter, and N1 (= N + K0) likelihoods. The data is input to the decoder 2e. The decoder 2e performs turbo decoding on the N1 likelihood data (y, a) and outputs an information bit estimation result.

As described above, decoding errors can be reduced by performing a process of appropriately inserting and deleting dummy bits on the transmission side and the reception side.
By the way, when the above method is applied to a radio base station including the transmission processing unit shown in FIG. 41, it becomes a problem how to insert and delete dummy bits.
In particular, the coding apparatus is configured in consideration of dummy bit insertion / deletion positions, coding rate constant / variable, code block division, size after dummy bit insertion, etc. There is a need to.
Further, it is necessary to insert dummy bits into the information bits so that decoding errors are effectively reduced.
In addition, in the case of a code that performs interleaving and deinterleaving, for example, a turbo code, when encoding, it is necessary to insert dummy bits into information bits so as to effectively reduce decoding errors in consideration of interleaving and deinterleaving.

PCT / JP2005 / 367 Paragraph 0104 of JP-T 2004-531972 (JP2004-531972)

3GPP, TS25.212v5.9.0

From the above, the object of the present invention is to consider the insertion / deletion position of dummy bits, whether the coding rate is constant / variable, whether code blocks are divided, or the size after dummy bit insertion. Thus, various transmission methods are provided.
Another object of the present invention is to insert dummy bits into information bits so that decoding errors can be effectively reduced.
Another object of the present invention is to provide a dummy bit at an information bit position that effectively reduces decoding errors in consideration of interleaving and deinterleaving in the case of a code that performs interleaving and deinterleaving, for example, a turbo code. Is to insert.
Another object of the present invention is to insert dummy bits into information bits so that decoding errors can be reduced and the coding rate becomes a required value.

The present invention systematically encodes information bits in which dummy bits are inserted, transmits a systematic code formed by deleting the dummy bits, and inserts the dummy bits deleted on the receiving side into the received systematic code. It is the transmission method in the communication system to decode.
According to a first transmission method of the present invention, a step of determining a size of a dummy bit to be inserted into an information bit based on a specified coding rate, a transmission rate, or a bit length that determines the transmission rate, the information bit and the dummy bit A step of dividing the information bits into the number of code blocks when the total size of the bits is larger than a prescribed size, a step of inserting dummy bits into each divided information bit in a predetermined pattern, and information bits into which the dummy bits are inserted In addition to systematic coding, the system includes a step of generating a systematic code by deleting the dummy bits from the systematic bits, and a step of transmitting the systematic code.
According to a first transmission method of the present invention, a step of determining a size of a dummy bit to be inserted into an information bit based on a specified coding rate, a transmission rate, or a bit length that determines the transmission rate, Inserting the information bit and the dummy bit when the total size of the information bit and the dummy bit is larger than the prescribed size, dividing the information bit into which the dummy bit is inserted, systematically coding the divided information bit, A step of generating a systematic code by deleting bits, and a step of transmitting the systematic code.

According to the present invention, a dummy bit is inserted into an information bit, a parity bit created from the information bit is added to the information bit for turbo coding, and the dummy bit is deleted from the turbo code. It is possible to reduce decoding errors by transmitting a code, receiving the systematic code on the receiving side, and inserting the dummy bit deleted on the transmitting side with maximum likelihood into the received systematic code and performing turbo decoding. .
Also, according to the present invention, dummy bits can be easily deleted from systematic bits by providing a dummy bit deletion unit in the physical layer HARQ function unit or the code unit. Further, according to the present invention, it is possible to perform processing for deleting dummy bits from systematic bits simultaneously with puncturing processing for parity bits of systematic codes. Therefore, dummy bit deletion can be performed without affecting the total transmission processing time.
According to the present invention, it is possible to improve the decoding characteristics because the dummy bits are distributed and inserted into the information bits by setting the continuous length of the dummy bits below the set value. Also, since the dummy bits are dispersed and inserted into the information bits excluding the peripheral portion at the beginning or end of the information bits, the decoding characteristics can be improved. Also, when a code that requires interleaving is employed, the decoding characteristics can be improved because the dummy bits are inserted in a distributed manner by excluding the bit positions that follow the information after interleaving.
According to the present invention, when a turbo code is employed, the pattern of dummy bit insertion positions is made as uniform as possible in both the first and second inputs, which are the inputs of the first element encoder and the second element encoder. As a result, the decoding characteristics can be improved.

FIG. 3 is a configuration diagram of a transmission processing unit of a radio base station according to the first embodiment. FIG. 6 is an explanatory diagram of dummy bit insertion processing according to the first embodiment. FIG. 6 is an explanatory diagram of dummy bit deletion processing according to the first embodiment. FIG. 3 is a configuration diagram of a code block dividing unit according to the first embodiment. It is a block diagram of a turbo code | symbol part. FIG. 10 is an explanatory diagram of dummy bit insertion processing of the radio base station according to the second embodiment. It is a principal part block diagram of 2nd Example. It is a dummy bit insertion processing flow of the second embodiment. It is explanatory drawing of the dummy bit insertion method after the code division | segmentation of 2nd Example. It is a dummy bit insertion processing flow of the second embodiment. It is explanatory drawing of the insertion method which inserts a dummy bit before code | symbol division | segmentation. It is explanatory drawing of the dummy bit insertion process of 3rd Example. It is a dummy bit insertion processing flow of the third embodiment. It is explanatory drawing of the dummy bit insertion process of 4th Example. It is a dummy bit insertion processing flow of the fourth embodiment. It is explanatory drawing of the dummy bit insertion process of 5th Example. It is a dummy bit insertion processing flow of the fifth embodiment. It is explanatory drawing of dummy bit insertion of 6th Example. It is a principal part block diagram of the transmission process part of 6th Example. It is a dummy bit insertion processing flow of the sixth embodiment. It is dummy bit insertion explanatory drawing of 7th Example. It is a principal part block diagram of the transmission process part of 7th Example. It is a dummy bit insertion processing flow of the seventh embodiment. In the seventh embodiment, dummy bit values are random patterns. It is an example of a dummy bit insertion pattern in the eighth embodiment. It is explanatory drawing of a dummy bit insertion position. It is explanatory drawing of the dummy bit insertion position in consideration of interleaving. The required Eb / N0 characteristic (decoding characteristic) with respect to the coding rate. It is a block diagram of the transmission process part in the wireless base station of 9th Example. It is explanatory drawing of the insertion position pattern of the dummy bit of 10th Example. It is a flow of dummy bit position change in the tenth embodiment. It is explanatory drawing of the dummy bit position change in 10th Example. It is a block diagram of the turbo encoder of 11th Example. It is a block diagram of the turbo decoding part of 11th Example. It is organization code explanatory drawing. It is a block diagram of the conventional communication system which block-encodes in a transmitter and decodes in a receiver. It is a block diagram of a turbo code | symbol part. It is a block diagram of a turbo decoding part. 1 is a configuration diagram of a 3GPP W-CDMA mobile communication system. FIG. It is a common channel explanatory drawing in HSDPA. It is a block diagram of a transmission processing unit of a 3GPP W-CDMA radio base station. It is a data format explaining a transmission process. It is explanatory drawing of the encoding / decoding method using a dummy bit. FIG. 44 is a configuration diagram of a communication system that realizes the encoding / decoding method of FIG. 43.

(A) First Embodiment FIG. 1 is a block diagram of a transmission processing unit 30 of a radio base station according to a first embodiment. Components identical with those of the conventional transmission processing unit of FIG. The transmission processing unit 30 transmits information (packets) to the mobile station using the HSDPA common channel HS-PDSCH.
The transmission processing unit 30 includes a CRC adding unit 21, a bit scrambling unit 22, a code block dividing unit 23, a dummy bit inserting unit 31, a channel coding unit (encoding unit) 24, a physical layer HARQ function unit 25, and a physical channel separating unit 26. HS-DSCH interleaving section 27, constellation rearrangement section 28, physical channel mapping section 29, and transmission section 32 for transmitting information. The dummy bit insertion unit 31 is provided between the code block division unit 23 and the encoding unit 24, and inserts dummy bits into information bits.
The physical layer HARQ function unit 25 includes a bit division unit 25a, a first rate matching unit 25b, a second rate matching unit 25c, and a bit combination unit 25d. The first rate matching unit 25b includes a dummy bit deleting unit 25b-3 for deleting dummy bits from the systematic bits in addition to the rate matching processing units 25b-1 and 25b-2 for the parity 1 and 2, and a second rate matching unit 25c is provided with rate matching processing units 25c-1 and 25c-2 for parity 1 and 2 and a rate matching processing unit 25c-3 for systematic bits as in the conventional example.
The dummy bit deleting unit 25b-3 deletes the dummy bit bits inserted into the systematic bits by the dummy bit inserting unit 31. Conventionally, the first rate matching unit 25b passes through the systematic bits without performing any processing. In the first embodiment, the punctures for the parity bits 1 and 2 by the rate matching processing units 25b-1 and 25b-2 are performed. Simultaneously with the ring process, the dummy bit deletion unit 25b-3 performs the dummy bit deletion process.

FIG. 2 is an explanatory diagram of dummy bit insertion processing, and FIG. 3 is an explanatory diagram of dummy bit deletion processing. The code block dividing unit 23 performs code block segmentation on the bit-scrambled data set D2. That is, the code bit division is performed by obtaining the dummy bit size K0 from the specified coding rate R and comparing the size K1 (= K + K0) of the information bit size K and the dummy bit size K0 with the specified size Z. The necessity determination of is performed. In the turbo code, Z ≦ 5114 because 40 ≦ K1 ≦ 5114.
If the dummy bit size is K0 and the information bit size is K, the coding rate R when the dummy bit is deleted and transmitted is a turbo code.
R = K / {K + 2 (K + K0)} (1)
Therefore, the dummy bit size K0 is
K0 = (K-3KR) / 2R (2)
It is obtained as When the total size K1 (= K + K0) exceeds the specified size Z, the code block dividing unit 23 determines the number C of code blocks and the code block size, divides the data set D2, and divides C sets (C = The code blocks 1 and 2 in 2) are made into a plurality of code blocks having the same data size ((a) in FIG. 2). If the data is not divisible by the number of divisions, adjust by inserting filler bits. The value of the filler bit is set to 0 and inserted at the beginning of the original data.
The dummy bit insertion unit 31 inserts dummy bits of size K0 / 2 into each code block ((b) of FIG. 2), and the encoding unit 24 encodes each code block in which the dummy bits are inserted, For example, turbo coding is performed ((c) in FIG. 2).
The bit division unit 25a of the physical layer HARQ function unit 25 converts the code of each code block output from the encoding unit 24 into (1) system bit + dummy bit, (2) parity bit 1, and (3) parity bit 2. Each is divided and serially connected to each other (see (a) of FIG. 3). Next, the first rate matching unit 25b of the physical layer HARQ function unit 25 checks whether the total bit length of the data set D5 is larger than the specified buffer size NIR, and when it is larger, the parity 1 and parity 2 are set so that the size becomes NIR. Is punctured, and at the same time, dummy bits are deleted from the systematic bits (FIG. 3 (b)).
Next, the second rate matching unit 25c of the physical layer HARQ function unit 25 performs rate matching on the data set D61 of systematic bits, parity 1 and 2 shown in FIG. 3B according to the designated H-ARQ transmission parameter ( Repeat or puncture). Thereafter, processing similar to that of the prior art is performed, and a systematic code formed by deleting dummy bits is transmitted. On the receiving side, the systematic code is received, and dummy bits deleted on the transmitting side are inserted into the received systematic code with maximum likelihood, and turbo decoding is performed to obtain information bits. In HSDPA, information necessary for reception (destination, modulation method, dummy bit size, dummy bit insertion method, etc.) is notified to the receiving apparatus in advance through the common channel HS-SCCH as necessary. Accordingly, since the dummy bit insertion position on the transmission side is known in the receiving apparatus, the dummy bit with the maximum likelihood is inserted into the position for decoding.
FIG. 4 is a block diagram of the code block dividing unit 23. The dummy bit size calculating unit 23a calculates the dummy bit size K0 based on the designated coding rate R, and the code block number / code block size determining unit 23b The number of code blocks and the code block size are determined based on the total size K1 (= K + K0) of the information bit size K and the dummy bit size K0 and the specified size Z, and the dividing unit 23c sets the bit-scrambled data set. D2 is divided into the designated number of divisions, the dummy bit insertion unit 31 inserts dummy bits of size K0 / 2 into each code block, and the encoding unit 24 applies to each code block with the dummy bits inserted, respectively. Turbo coding is performed.
As described above, according to the first embodiment, dummy bits are inserted into information bits, a parity bit created from the information bits is added to the information bits, turbo-coded, and the dummy bits are deleted from the turbo code. The decoding code is reduced by receiving the systematic code, receiving the systematic code on the receiving side, and inserting the dummy bit deleted on the transmitting side with maximum likelihood into the received systematic code and turbo decoding. can do.
Further, according to the first embodiment, since the dummy bit deleting unit 25b-3 has already been separated into systematic bits, the dummy bits can be easily deleted from the systematic bits. Further, since dummy bit deletion processing can be performed simultaneously with puncturing processing for parity bits 1 and 2, dummy bit deletion does not affect the total transmission processing time.

Modification Example Although the dummy bit deletion unit 25b-3 of the physical layer HARQ function unit 25 has deleted dummy bits as described above, it can be deleted in the turbo coding unit. FIG. 5 is a block diagram of the turbo encoder 24, where 24a is a first element encoder that encodes information bits with dummy bits inserted, 24b is an interleaver that interleaves information bits with dummy bits inserted, and 24c. Is a second element encoder that encodes the interleave result, 24d is a dummy bit deletion section that deletes dummy bits, and 24e converts the outputs of the element encoders 24a and 24b and the dummy bit deletion section 24c into serial data and outputs them. P / S conversion unit. If the dummy bits are deleted in the turbo code part as described above, the dummy bits can be easily deleted from the systematic bits because they are already separated into systematic bits.

(B) Second Embodiment In the second embodiment, encoding is performed by inserting dummy bits so that the coding rate becomes a fixed value, and the entire bit size of transmission data is made constant at Ndata. Ndata is the number of codes × physical channel size.
FIG. 6 is an explanatory diagram of dummy bit processing of the radio base station of the second embodiment, and the configuration of the transmission processing unit has the same configuration as that of the first embodiment of FIG.
The code block segmentation unit 23 performs code block segmentation on the data set D2 that has been bit-scrambled as in the first embodiment. That is, the size K0 of the dummy bit is obtained so that the specified coding rate R is obtained, and the total size K1 (= K + K0) of the size K of the information bit and the size K0 of the dummy bit is compared with the size comparison result of the size Z Then, it is determined whether or not code block division is necessary, and code block division is performed ((a) of FIG. 6).
The dummy bit insertion unit 31 inserts dummy bits of size K0 / 2 into each code block ((b) of FIG. 6), and the encoding unit 24 encodes each code block with the dummy bits inserted, For example, turbo encoding is performed ((c) in FIG. 6).
The first rate matching unit 25b of the physical layer HARQ function unit 25 checks whether the bit length of the code is larger than the specified buffer size NIR, and when it is larger, the puncturing is performed on the parity 1 and parity 2 so that the size becomes NIR. At the same time, the dummy bits are deleted from the systematic bits. Next, the second rate matching unit 25c of the physical layer HARQ function unit 25 performs rate matching (repetition or puncturing) so that the code length is Ndata ((d) in FIG. 6).
Thereafter, the same processing as in the prior art is performed to transmit a systematic code not including dummy bits. On the receiving side, the systematic code is received, and dummy bits deleted on the transmitting side are inserted into the received systematic code with maximum likelihood, and turbo decoding is performed to obtain information bits.

FIG. 7 is a block diagram of the main part of the transmission processing unit of the second embodiment, and the same parts as those of the first embodiment of FIG. The difference is that a physical layer HARQ function unit 25 having a dummy bit deletion unit and a second rate matching unit is added.
FIG. 8 is a flowchart of dummy bit insertion processing according to the second embodiment. The dummy bit size K0 is determined so as to obtain the prescribed coding rate R (step 501), and the total size K1 (= K + K0) of the information bit size K and the dummy bit size K0 is calculated (step 502). ) Compare the total size K1 with the specified size Z (step 503). If K1 ≦ Z, insert a dummy bit of size K0 into the information bit without dividing the code block (step 504), and insert the dummy bit The process ends. On the other hand, if K1> Z, the number of code blocks / code block size is determined and code blocks are divided (step 505). Next, filler bits are inserted (step 506), dummy bits of size K0 / C (C is the number of divisions and K0 / 2 if C = 2) are inserted into each code block (step 507), and dummy bits are inserted. The process ends.
FIG. 9 is an explanatory diagram of a dummy bit insertion method after code division. When inserting dummy bits, the same number of dummy bits are uniformly allocated for each code block, and the dummy bit insertion position and the value (0, 1) of the dummy bits are made the same.
Although FIG. 8 shows a case where dummy bits are inserted after code block division, dummy bits are inserted before code block division, and when division is necessary, the dummy bits are evenly distributed among the code blocks. Code blocks can also be divided.
FIG. 10 is a flowchart of dummy bit insertion processing according to the second embodiment, in which dummy bit insertion processing (step 511) is arranged before processing step 503 for comparing the size of the total size K1 and the specified size Z. . FIG. 11 is an explanatory diagram of an insertion method for inserting dummy bits before code division. When code blocks are divided, the dummy bit distribution is uniform in each code block and the dummy bit insertion positions are uniform. To be.
As described above, according to the second embodiment, the dummy bit size is determined so that the required coding rate is obtained, and the rate matching is performed so that Ndata given by the H-ARQ transmission parameter is transmitted. it can. In addition, decoding characteristics can be improved by inserting dummy bits uniformly.
The dummy bit insertion method shown in FIGS. 9 and 11 is not limited to the second embodiment, but can be applied to all the embodiments.

(C) Third Embodiment The third embodiment is an example in which dummy bits are inserted so that the total bit length of the code is equal to Ndata. FIG. 12 is an explanatory diagram of the dummy bit insertion process of the third embodiment, FIG. 13 is a dummy bit insertion process flow, and the configuration of the transmission processing unit is the same as that of the first embodiment of FIG.
The code block dividing unit 23 calculates the size K0 of the dummy bits to be inserted so that the total bit length is equal to Ndata (step 551). A code size when a dummy bit of size K0 is inserted into information bits of size K to perform turbo coding, and the dummy bit is deleted and transmitted is K + 2 (K + K0). Therefore, the following formula
Ndata = K + 2 (K + K0) (3)
And the dummy bit size K0 is
K0 = (Ndata−3K) / 2 (4)
It is.
Next, the size of the information bit with the dummy bit inserted, K1 = K + K0, is compared with the specified size Z (= 5114) (step 552). If K1 ≦ Z, the code block is not divided and the dummy bit of size K0 Is inserted into the information bits (step 553), and the dummy bit insertion process is terminated. On the other hand, if K1> Z, the number of code blocks / code block size is determined and code blocks are divided (FIG. 12 (a), step 554). Next, filler bits are inserted (step 555), and dummy bits of size K0 / C (C is the number of divisions and K0 / 2 if C = 2) are inserted into each code block ((b), FIG. 12). Step 556), the dummy bit insertion process is terminated.
The encoding unit 24 performs encoding, for example, turbo encoding, on each code block in which dummy bits are inserted ((c) in FIG. 12, step 557). Further, the physical layer HARQ function unit 25 deletes the dummy bits from the systematic bits ((d) in FIG. 12, step 558). Since the code length after deleting the dummy bits is equal to Ndata, the physical layer HARQ function unit 25 does not perform rate matching (repetition or puncturing).
Thereafter, processing similar to that of the prior art is performed, and a systematic code formed by deleting dummy bits is transmitted. On the receiving side, the systematic code is received, and dummy bits deleted on the transmitting side are inserted into the received systematic code with maximum likelihood, and turbo decoding is performed to obtain information bits.
According to the third embodiment, transmission can be performed by inserting dummy bits such that the coding rate R (= K / Ndata) is variable and the code length is equal to Ndata.

(D) Fourth Embodiment The fourth embodiment is an embodiment when code block division is not performed (the number of code blocks = 1). FIG. 14 is an explanatory diagram of dummy bit insertion processing of the fourth embodiment, FIG. 15 is a dummy bit insertion processing flow, and the configuration of the transmission processing unit is the same as that of the first embodiment of FIG.
If the size K1 (= K + K0) of the dummy bit size K0 and the information bit size K determined from the specified coding rate exceeds the specified size Z, the total size K1 is the specified size Z in the fourth embodiment. Adjust the dummy bit size so that
The code block dividing unit 23 determines the dummy bit size K0 by the equation (2) so as to obtain the prescribed coding rate R (step 601), and the total size K1 of the information bit size K and the dummy bit size K0. (= K + K0) is calculated (step 602), and the total size K1 is compared with the specified size Z (step 603).
If K1 ≦ Z, a dummy bit of size K0 is inserted into an information bit of size K (FIG. 14 (a), step 604). On the other hand, if K1> Z, the amount ΔK exceeding the specified size Z is expressed by the following equation: ΔK = K1−Z (5)
The dummy bit size K0 is calculated by
K0 = K0−ΔZ
(Step 605). Thereafter, dummy bits of size K0 are inserted into information bits of size K (FIG. 14 (a), step 604).
As described above, when the dummy bit insertion process is completed, the encoding unit 24 performs encoding, for example, turbo encoding, on the code block in which the dummy bits are inserted ((b) in FIG. 14, step 606). Further, the physical layer HARQ function unit 25 deletes the dummy bits from the systematic bits and performs rate matching processing so that the code length becomes equal to Ndata ((c) of FIG. 14, step 607).
Thereafter, the same processing as in the prior art is performed to transmit a systematic code not including dummy bits. On the receiving side, the systematic code is received, and dummy bits deleted on the transmitting side are inserted into the received systematic code with maximum likelihood, and turbo decoding is performed to obtain information bits.
According to the fourth embodiment, even when code block division is not performed, the maximum number of dummy bits can be inserted and the code length can be set to Ndata for transmission. For this reason, the dummy insertion effect can be enhanced when the code block is not divided.

(E) Fifth Embodiment The fifth embodiment is an embodiment in which the size of the dummy bit is determined so that the total size of the dummy bit and the information bit becomes the specified size Z in each code block when the code block is divided. FIG. 16 is an explanatory diagram of the dummy bit insertion process of the fifth embodiment, FIG. 17 is a dummy bit insertion process flow, and the configuration of the transmission processing unit is the same as that of the first embodiment of FIG.
The code block dividing unit 23 determines the dummy bit size K0 by the equation (2) so as to obtain the prescribed coding rate R (step 651), and the total size K1 of the information bit size K and the dummy bit size K0. (= K + K0) is calculated (step 652), and the total size K1 is compared with the specified size Z (step 653).
If K1 ≦ Z, a dummy bit of size K0 is inserted into an information bit of size K (step 654). Note that. Dummy bits can be inserted so that the code block size becomes a specified size.
On the other hand, if K1> Z, the number of code blocks / code block size is determined, and code blocks are divided (FIG. 16 (a), step 655). Next, filler bits are inserted (step 656), dummy bits are inserted in each code block so that the size is the prescribed size Z (step (b) in FIG. 16, step 657), and the dummy bit insertion process is terminated.
The encoding unit 24 performs, for example, turbo encoding on each code block in which dummy bits are inserted ((c) of FIG. 16). The physical layer HARQ function unit 25 deletes dummy bits from the systematic bits and performs rate matching processing so that the code length is equal to Ndata.
Thereafter, the same processing as in the prior art is performed to transmit a systematic code not including dummy bits. On the receiving side, the systematic code is received, and dummy bits deleted on the transmitting side are inserted into the received systematic code with maximum likelihood, and turbo decoding is performed to obtain information bits.
According to the fifth embodiment, the dummy bits can be inserted and encoded so that the total size of the dummy bits and the information bits becomes the specified size Z in each code block, and the dummy bits can be deleted and transmitted. In this case, since the dummy bit size to be inserted can be increased, the dummy insertion effect can be enhanced.

(F) Sixth Embodiment The sixth embodiment is an embodiment in which dummy bits are inserted before bit scrambling, FIG. 18 is an explanatory diagram of dummy bit insertion, FIG. 19 is a block diagram of the main part of the transmission processing section, and FIG. Reference numeral 20 denotes a dummy bit insertion processing flow.
The dummy bit size calculation unit 31a of the dummy bit insertion unit 31 determines the dummy bit size K0 by the equation (2) so that the prescribed coding rate R is obtained (step 701), and the information bit size K and the dummy bit are determined. The total size K1 (= K + K0) with the size K0 is calculated (step 702), and the dummy bit insertion unit 31b adds all 0 to the information bits ((a) in FIG. 18) added with CRC by the CRC adding unit 21. Are inserted ((b) of FIG. 18, step 703). The dummy bits are not limited to all zeros.
Next, the bit scrambler 22 bit-scrambles the information bit with the dummy bit inserted and inputs it to the code block divider 23 ((c) in FIG. 18, step 704).
The code block number / code block size determining unit 23b of the code block dividing unit 23 compares the size of the bit scrambled data set D2 (total size of information bits and dummy bits) K1 with the size of the specified size Z (step 705). .
If K1 ≦ Z, code division is not performed. On the other hand, if K1> Z, the number of code blocks / code block size is determined, and the division unit 23c performs code block division (step 706). Next, a filler bit is inserted (step 707).
Thereafter, as in the first embodiment, the encoding unit 24 performs turbo encoding on each code block in which dummy bits are inserted, and the physical layer HARQ function unit 25 deletes dummy bits and performs predetermined rate matching. Processing is performed and a systematic code not including dummy bits is transmitted. On the receiving side, the systematic code is received, and dummy bits deleted on the transmitting side are inserted into the received systematic code with maximum likelihood, and turbo decoding is performed to obtain information bits.
According to the sixth embodiment, dummy bits can be inserted before bit scrambling.

(G) Seventh Embodiment The seventh embodiment is an embodiment in which dummy bits are inserted after bit scrambling, FIG. 21 is an explanatory diagram of dummy bit insertion, FIG. 22 is a principal block diagram of a transmission processing unit, and FIG. It is a dummy bit insertion processing flow.
The bit scrambler 22 adds a bit scramble process to the information bits ((a) in FIG. 21) added with the CRC by the CRC adder 21 ((b) in FIG. 21, step 751). Next, the dummy bit size determination unit 31a of the dummy bit insertion unit determines the dummy bit size K0 by the expression (2) so that the prescribed coding rate R is obtained (step 752), and the information bit size K and the dummy bit are determined. The total size K1 (= K + K0) with the bit size K0 is calculated (step 753), and the dummy bit insertion unit 31b inserts all 1 dummy bits of size K0 into the bit-scrambled information bits (FIG. 21). (C), step 754). Note that all 0s are inappropriate as dummy bits.
The code block number / code block size determining unit 23b of the code block dividing unit 23 compares the total size K1 of information bits and dummy bits with the size of the specified size Z (step 755). If K1 ≦ Z, code division is not performed. On the other hand, if K1> Z, the number of code blocks / code block size is determined, and the division unit 23c performs code block division (step 756). Next, a filler bit is inserted (step 757).
Thereafter, as in the first embodiment, the encoding unit 24 performs turbo encoding on each code block in which dummy bits are inserted, and the physical layer HARQ function unit 25 deletes dummy bits and performs predetermined rate matching. Processing is performed, and a systematic code not including dummy bits is transmitted. On the receiving side, the systematic code is received, and dummy bits deleted on the transmitting side are inserted into the received systematic code with maximum likelihood, and turbo decoding is performed to obtain information bits.
In the above example, the dummy bit insertion unit 31 has inserted all 1 dummy bits, but the dummy bit value can be a random pattern as shown in FIG.
According to the seventh embodiment, dummy bits can be inserted after bit scrambling.

(H) Eighth Embodiment The eighth embodiment is an embodiment of an insertion pattern of dummy bits into information bits. The pattern in which systematic bits and dummy bits are alternately arranged as an insertion pattern as shown in FIG. 25A can improve the decoding characteristics as compared with a pattern in which dummy bits are arranged in a single place before and after the information bits. .
However, the pattern of the alternating arrangement is when the sizes of the systematic bit and the dummy bit are the same, and cannot be arranged alternately if they are different. Therefore, the dummy bits are inserted into the systematic bits so as to allow the dummy bits to continue for the designated length. Thus, even if the dummy bits are arranged in a distributed manner by setting the continuous length of the dummy bits to be equal to or less than the set value, the decoding characteristics (decoding error characteristics) can be improved. For example, when the information bits and dummy bits have the same size and the continuous length is 2, as shown in FIG. 15B, 2 information bits and 2 dummy bits are alternately arranged. When the continuous length is 3, as shown in FIG. 15C, 3 information bits and 3 dummy bits are alternately arranged.
Further, as shown in FIG. 26, dummy bits are not inserted into the surrounding STA and TLA at the beginning or end of information. This is because in Viterbi decoding and MAP decoding, the reliability of codes at the beginning and end of information is sufficiently high. Therefore, as shown in FIG. 26, dummy bits are distributedly inserted into the area excluding the peripheral STA and TLA at the beginning or end of the information.
Also, as shown in FIG. 27, the bit positions A1 to A4 that come to the specified number of positions at the beginning and the back of the data after interleaving are specified in advance from the internal interleaving pattern of the turbo code. Also, dummy bits are not inserted into these positions A1 to A4. The reason is the same as in FIG.

(I) Ninth embodiment
The 3GPP turbo code has a characteristic that when the coding rate becomes a specific value by puncturing, the characteristic degradation is locally larger than the surrounding coding rates. FIG. 28 is an explanatory diagram showing such characteristic degradation, where A is the decoding characteristic without dummy bit insertion, the horizontal axis is the coding rate, and the vertical axis is the required Eb / No for obtaining a predetermined error rate. is there. As is apparent from this decoding characteristic, when the coding rate reaches a specific value (7/11, 7/9, 7/8), the required Eb / No becomes larger than the surrounding coding rate and the characteristic deteriorates. . Therefore, in the ninth embodiment, it is monitored whether the coding rate after puncturing has become a value close to a specific value (a value in the specific range S1 to S3), and when it becomes a value in the specific range S1 to S3, Dummy bits are inserted before puncturing, and the decoding characteristic is shifted as indicated by B, so that the coding rate becomes a value outside the specific range S1 ′ to S3 ′ determined by the decoding characteristic B. To prevent. The amount of dummy bit insertion is determined so that it deviates from the region at the peak of the characteristic B peak after the shift.
FIG. 29 is a configuration diagram of the transmission processing unit in the radio base station of the ninth embodiment. The transmission processing unit 30 includes a CRC adding unit 21, a bit scramble unit 22, a code block dividing unit 23, a dummy bit insertion control unit 41, Channel coding unit (encoding unit) 24, physical layer HARQ function unit 25, physical channel separation unit 26, HS-DSCH interleaving unit 27, constellation rearrangement unit 28, physical channel mapping unit 29, transmission unit (not shown) It has.
The dummy bit insertion control unit 41 is provided between the code block dividing unit 23 and the encoding unit 24, and controls whether dummy bits are inserted into information bits based on the coding rate. That is, the dummy bit insertion control unit 41 calculates the coding rate R in consideration of puncturing in the physical layer HARQ function unit 25 (step 801). If the information bit length is K, the parity bit length of the systematic code obtained by encoding the information is M, and the number of puncturing bits is P, the coding rate R is
R = K / (K + MP)
It becomes. In the case of a turbo code, since M = 2K, R = K / (3K-P).
The dummy bit insertion control unit 41 checks whether or not the calculated coding rate R has become a value in the range S1 to S3 of ± Δ centered on the specific values 7/11, 7/9, and 7/8 (step 802). ). If the value is not within the range, the dummy bit insertion control unit 41 does not insert the dummy bit. If the value is within the range, the decoding characteristic is shifted from A to B and the coding rate is within the specific range S1. A dummy bit is inserted into the information bit so as to be a value outside '-S3' (step 803).
According to the ninth embodiment, since the dummy bits are inserted so that the coding rate does not become a value within a specific range that deteriorates the decoding characteristic, it is possible to prevent the deterioration of the decoding characteristic.

(J) Tenth Embodiment When a turbo code is used as a code, input bits (referred to as “first input” and “second input”, respectively) of the first element encoder and the second element encoder of the turbo encoder section. In both cases, decoding characteristics can be improved by making the pattern of dummy bit insertion positions as uniform as possible.
For this reason, for the several bits before and after the dummy bit insertion position, different dummy bits are avoided as much as possible. That is, when the number of information bits K and the number of dummy bits K0 are K0 ≦ K, the dummy bits are not adjacent to each other in both the first input and the second input, and the dummy bits of the first and second inputs An arrangement where the insertion positions are equal is an ideal arrangement. When K0> K, the ideal arrangement is that the information bits are not adjacent to each other in both the first input and the second input, and the positions of the information bits of the first and second inputs are equal. . When K0> K and there are more dummy bits than information bits, in principle, at least two dummy bits are adjacent to each other. In this case, the uniformity is achieved by switching the positions of dummy bits and information bits.
Only when the ratio of K0 and K is not an integer ratio and when such an arrangement is impossible due to the interleaved position relationship, the replacement of the dummy bit position and the information bit position is permitted. However, this replacement is performed almost equally for each of the first and second inputs.
For example, as shown in FIG. 30A, when K = K0, a completely uniform arrangement is performed with respect to the first input (alternate arrangement), and the interleave pattern arrangement P is applied to generate the second input. . In the second input, a portion where the burst length (continuous length) of the dummy bit is 3 or more is obtained, and if present, the dummy bit position d where the burst length becomes 1 or 2 when a certain dummy bit is changed to an information bit is set. Ask. Next, the position at the first input corresponding to the dummy bit position d is obtained by Q (d). However, Q is a deinterleave pattern arrangement, and P (Q (d)) = d. In the first input, the position Q (d) ± 1 on both sides of the position Q (d) is currently an information bit, but the bit position (P (Q (d) +1) of the second input corresponding to these positions. , P (Q (d) -1)), the bit position with the shorter burst length of the generated dummy bit is selected (P (Q (d) +1 in the figure). 30B, the dummy bit at position Q (d) in the first input and the information bit at position Q (d) +1 in the first input are exchanged, that is, position Q (d) in the first input. Is changed from the dummy bit to the information bit, and the position Q (d) +1 is changed from the information bit to the dummy bit, so that the continuous length of the dummy bit at the second input after interleaving is 2 or less. be able to.

FIG. 31 is a flow of an efficient algorithm for changing the dummy bit position so as to satisfy the conditions described in FIG. The input information bit size K, the dummy bit size K0, and the combined bit size K1 are K1 = K + K0. Also, let P (i) and Q (i) be an interleave pattern array and vice versa. That is, Q (P (i)) = i. Further, the number of dummy bits whose position is determined is Nd, and the threshold Th for position determination is Th = 10. Also, each bit is associated with a weight variable W (i) as shown in FIG.
First, the counter is initialized to Nd = 0, and the total weight coefficient W (i) is initialized to 0 (step 901).
Next, the following is repeated for i = 0 to K1-1. That is, if i = 0 and i <K1 (steps 902 to 903), it is checked whether W (i) ≦ Th (step 904). If W (i) ≦ Th, the position i is set as a dummy bit position (step 905), and the weight variable is updated as follows (step 906).
W (i) = 300
W (i + 1) + 50 = W (i + 1) W (i-1) + 50 = W (i-1)
W (i + 2) + 10 = W (i + 2) W (i-2) + 10 = W (i-2)
W (Q (P (i) +1)) + 50 = W (Q (P (i) +1)) W (Q (P (i) -1)) + 50 = W (Q (P (i) -1))
W (Q (P (i) +2)) + 10 = W (Q (P (i) +2)) W (Q (P (i) -2)) + 10 = W (Q (P (i) -2))
However, no processing is performed when x <0, x ≧ K1 with respect to W (x).
Next, the position determination dummy bit number Nd is counted up (Nd + 1 = Nd, step 907), and whether or not Nd <K0 is checked (step 908). If so, i is incremented (step 909), and the processing after step 903 is continued. If W (i)> Th in step 904, i is incremented immediately (step 909), and the processing after step 903 is continued.
On the other hand, if i = K1 in step 903, it is checked whether Nd <K0 (step 910). If Nd ≧ K0, the process ends. If Nd <K0, Wmin is set to W (i). The minimum value is set (step 911), then Th = Wmin + 20 is set (step 912), and the processing after step 902 is repeated.
The method of uniformly inserting dummy bits with respect to input information bits, which has been the basis of the embodiments so far, may cause inconvenient situations. For example, when the turbo code is adopted, the input of the second element encoder of the turbo code unit becomes a pattern after interleaving. For this reason, if the dummy bits are simply inserted uniformly with respect to the information bits before the interleaving, the positions of the dummy bits are changed by the interleaving, and the dummy bit position of the second input of the second element encoder is changed. It is not uniform. As a result, an undesired pattern (continuous long dummy bits) is generated at the second input, which causes deterioration of decoding characteristics. Therefore, in the tenth embodiment, the dummy bit insertion position is determined by the above algorithm so that the continuous length of the dummy bits does not become long. That is, the dummy bit insertion position is sequentially determined, and the weights immediately adjacent to the determined dummy bit position and the next adjacent weight are increased in both the first and second inputs to obtain the dummy bit position. The selection is made difficult so that the continuous length of the dummy bits is not increased.
Note that the present invention is not limited to the above algorithm, and any algorithm that makes the pattern of dummy bit insertion positions as uniform as possible in both the first and second inputs can be adopted in the tenth embodiment.
Although the method for determining the dummy bit size K0 has not been described above, as described in the second embodiment, the size of the dummy bit to be inserted into the information bits is determined based on the designated coding rate. Alternatively, as described in the third embodiment, the size K0 of the dummy bit to be inserted is calculated so that the code size becomes equal to the bit length Ndata determined from the transmission rate of the physical channel.

(K) Eleventh Embodiment In many cases, it is characteristically arranged that the dummy bit insertion positions are varied as much as possible at both the input of the first element code and the second element code of the turbo code section. It has proven to be advantageous. In order to realize this, the method of the tenth embodiment has a problem that the position generation algorithm is complicated, and the processing amount and processing time are increased. In the eleventh embodiment, the dummy bit insertion positions are easily varied as much as possible in both the first and second inputs of the first and second element codes.
FIG. 33 is a block diagram of the turbo encoder of the eleventh embodiment. The first element encoder 24a encodes information bits with dummy bits inserted, and the interleave unit 24b interleaves the information bits with dummy bits inserted. The second element encoder 24c encodes the interleave result, and the P / S converter 24e converts the output xb, xc and information bit xa of each element encoder 24a, 24b into serial data and outputs the serial data. The first and second dummy bit insertion units 51 and 52 insert dummy bits into the first and second inputs which are the inputs of the first and second element encoders 24a and 24b. Preferably, the first and second inputs are inserted so that they are dispersed as much as possible and as uniform as possible.
As described in the second embodiment, the dummy bit size K0 to be inserted has a code length of physical according to the expression (2) based on the specified coding rate or as described in the third embodiment. It is calculated by equation (4) so as to be equal to the bit length Ndata determined from the channel transmission rate.
In the configuration of FIG. 33, two element encoders are provided, but the first and second element code processes can be performed by one element encoder.
FIG. 34 is a block diagram of a turbo decoding unit on the receiving side that decodes the turbo code encoded by the encoder of FIG.
The first element decoder 61 performs decoding by using ya and yb among the received signals ya, yb and yc. The first element decoder 61 is an element decoder with a soft decision output, and outputs the likelihood of the decoding result. The first dummy bit deleting unit 62 deletes the first dummy bit from the decoding result of the first element decoder 61, and the interleaving unit 63 interleaves the decoding result from which the dummy bit is deleted, and inserts the second dummy bit. The unit 64 inserts the second dummy bit having the maximum likelihood into the interleaved decoding result.
The second element decoder 65 performs decoding using the decoding result of the first element decoder 61 subjected to the interleaving and second dummy bit insertion processing and the received signal yc. The second element decoder 65 is also an element decoder having a soft decision output, and outputs the likelihood of the decoding result. The second dummy bit deletion unit 66 deletes the second dummy bit from the decoding result of the second element decoder 65, and the deinterleaving unit 67 deinterleaves the decoding result from which the dummy bit has been deleted, and the first dummy bit The bit insertion unit 68 inserts the first dummy bit having the maximum likelihood into the deinterleaved decoding result and inputs it to the first element decoder 61. The first element decoder 61 repeats the above MAP decoding process using the output signal of the first dummy bit insertion unit 68 instead of the received signal ya. By repeating the above decoding operation a predetermined number of times, the error rate of the decoding result is reduced. A MAP element decoder is used as the first and second element decoders in such a turbo element decoder.
The above process of deleting and adding the maximum likelihood reliability of dummy bit data is to select a trellis path limited from the value of the dummy bit in the likelihood calculation of the decoder. It is also possible to select the trellis path directly instead of deleting.
In the configuration of FIG. 34, two element decoders are provided, but the first and second element decoding processes can be performed by one element decoder. Similarly, the first and second dummy bit deletion processing and the first and second dummy bit insertion processing can be performed by one dummy bit deletion unit and one dummy bit insertion unit.
According to the eleventh embodiment, the insertion positions independent of each other can be taken at both the input of the first element code 24a and the second element code 24c, and in particular, a uniform pattern is selected for both. It becomes possible. Further, the dummy bit deletion unit is not necessary.

(L) Effects of the Invention As described above, according to the present invention, dummy bits are inserted into information bits, a parity bit created from the information bits is added to the information bits and turbo-encoded, and the dummy code is generated from the turbo code. By transmitting a systematic code in which bits are deleted, receiving the systematic code on the receiving side, inserting dummy bits deleted on the transmitting side with maximum likelihood into the received systematic code, and performing turbo decoding Decoding errors can be reduced.
Also, according to the present invention, dummy bits can be easily deleted from systematic bits by providing a dummy bit deletion unit in the physical layer HARQ function unit or the code unit. Further, according to the present invention, it is possible to perform processing for deleting dummy bits from systematic bits simultaneously with puncturing processing for parity bits of systematic codes. Therefore, dummy bit deletion can be performed without affecting the total transmission processing time.
According to the present invention, it is possible to determine the size of a dummy bit so that the required coding rate is obtained, and perform rate matching so as to be Ndata given by the H-ARQ transmission parameter for transmission. In addition, decoding characteristics can be improved by inserting dummy bits uniformly.
According to the present invention, transmission can be performed by inserting dummy bits such that the coding rate R is variable and the code length is equal to Ndata.
According to the present invention, even when code block division is not performed, the maximum number of dummy bits can be inserted and the code length can be transmitted as Ndata. For this reason, the dummy insertion effect can be enhanced in the case where code blocks are not divided.
According to the present invention, the dummy bits are inserted and encoded so that the total size of the dummy bits and the information bits becomes the prescribed size Z in each code block, and the dummy bits are deleted and transmitted. Since the size can be increased, the dummy insertion effect can be enhanced.
According to the present invention, dummy bits can be inserted before bit scrambling or after bit scrambling.
According to the present invention, it is possible to improve the decoding characteristics because the dummy bits are distributed and inserted into the information bits by setting the continuous length of the dummy bits below the set value. Also, since the dummy bits are dispersed and inserted into the information bits excluding the peripheral portion at the beginning or end of the information bits, the decoding characteristics can be improved. Also, when a code that requires interleaving is employed, the decoding characteristics can be improved because the dummy bits are inserted in a distributed manner by excluding the bit positions that follow the information after interleaving.
According to the present invention, since the dummy bits are inserted so that the coding rate does not become a specific value that degrades the decoding characteristic, it is possible to prevent the deterioration of the decoding characteristic.
According to the present invention, when a turbo code is employed, the pattern of dummy bit insertion positions is made as uniform as possible in both the first and second inputs, which are the inputs of the first element encoder and the second element encoder. As a result, the decoding characteristics can be improved.
According to the present invention, when the turbo code is adopted, the dummy bit insertion position can be determined without depending on each other at the first and second inputs of the first element encoder and the second element code. In the second input, the dummy bit insertion position pattern can be easily made uniform, and the decoding characteristic can be improved.
Further, according to the present invention, since a dummy bit is inserted in the turbo encoder to generate a parity bit and a turbo code can be output without inserting the dummy bit into the systematic bit, a dummy for deleting the dummy bit from the systematic bit can be obtained. The bit deletion part can be made unnecessary.

21 CRC adding section 22 Bit scrambling section 23 Code block dividing section 24 Channel coding section (encoding section)
25 Physical layer HARQ function unit 26 Physical channel separation unit 27 HS-DSCH interleaving unit 28 Constellation rearrangement unit 29 Physical channel mapping unit 30 Transmission processing unit 31 Dummy bit insertion unit 32 Transmission unit for transmitting information

Claims (4)

A communication system that systematically encodes information bits in which dummy bits are inserted, transmits a systematic code obtained by deleting the dummy bits, and inserts and decodes the dummy bits deleted on the receiving side into the received systematic code In the transmission method in
Based on the specified coding rate, transmission rate, or bit length that determines the transmission rate, determine the size of the dummy bits to be inserted into the information bits,
The information bits is divided into code blocks when the information bits and the total size of the dummy bits is greater than the specified size,
Insert dummy bits in a predetermined pattern into each divided information bit ,
Systematically encode the information bits with the dummy bits inserted, and delete the dummy bits from the systematic bits to generate a systematic code,
Transmitting the tissue code;
A transmission method characterized by the above. A communication system that systematically encodes information bits in which dummy bits are inserted, transmits a systematic code obtained by deleting the dummy bits, and inserts and decodes the dummy bits deleted on the receiving side into the received systematic code In the transmission method in
Based on the specified coding rate, transmission rate, or bit length that determines the transmission rate, determine the size of the dummy bits to be inserted into the information bits,
Insert dummy bits into the information bits,
When the total size of the information bits and dummy bits is larger than the specified size, the information bits with the dummy bits inserted are divided,
Systematically encode the divided information bits, delete dummy bits from systematic bits, and generate systematic codes.
Transmitting the tissue code;
A transmission method characterized by the above. When determining the size of the dummy bits based on the specified coding rate, rate matching processing is performed so that the total size of the systematic codes from which the dummy bits are deleted is equal to the bit length that determines the transmission rate.
The transmission method according to claim 1 or 2, characterized in that When the dummy bits are inserted, dummy bits are further added so that the total size of the divided information bits and dummy bits becomes the specified size.
The transmission method according to claim 1.

Images