CN102299770A - United coding method and system for long term evolution (LTE) uplink control information (UCI) - Google Patents

Abstract

The invention discloses a unified coding method and system for LTE uplink control information, and belongs to the technical field of long-term evolution channel coding. In this method, a two-stage puncturing of a basic TBCC with a code rate of 1/5 is used to complete the channel coding of UCI at different lengths. The first stage selects the generator, and the second stage deletes redundant codeword bits: when replacing the LTE block code, the generator required by the first stage can be selected according to the size of the UCI and the determined puncture pattern, and the second stage deletes The number and position of bits can be used to obtain a code word with a code length of 20 or 32. When replacing the TBCC code of LTE, it is enough to select the three generators of the original LTE in the first stage, and the codeword is not deleted in the second stage. The present invention can use a unified method to encode/decode the UCI on the PUCCH and PUSCH, avoiding the selection between multiple encoding and decoding methods at the user end and the base station, and simplifying the encoding/decoding of the system structure.

Description

Unified encoding method and system for LTE uplink control information

technical field

The invention relates to a unified encoding method and system for LTE uplink control information, and belongs to the technical field of Long Term Evolution (LTE) channel encoding.

Background technique

LTE is a long-term evolution project of Universal Mobile Telecommunications System (UMTS), and is currently the most influential Beyond 3G (B3G) system, which can provide higher data rate, lower delay and larger system capacity and coverage.

The uplink of LTE adopts Single-Carrier Frequency Division Multiple Access (Single-Carrier Frequency Division Multiple Access, SC-FDMA) as the basic transmission technology of the physical layer. SC-FDMA divides the transmission bandwidth into multiple parallel orthogonal subcarriers, and uses a cyclic prefix (Cyclic Prefix, CP) to maintain the orthogonality of subcarriers in frequency selective channels. According to the duration, CP can be divided into two types: regular CP and extended CP. Based on SC-FDMA transmission technology, LTE defines uplink transmission resources from two dimensions of time domain and frequency domain: in terms of time domain, the largest unit is a radio frame of 10 milliseconds (ms), and each radio frame is divided into ten 1ms subframe, each subframe is divided into 2 slots (slot), each slot contains 7 SC-FDMA symbols under the regular CP, and only contains 6 SC-FDMA symbols under the extended CP; from the frequency domain It is said that 12 subcarriers are used as a unit. The time-frequency structure of the LTE uplink can be described by a resource block (Resource Block, RB). One RB refers to 12 subcarriers in one time slot. The RB can be further divided into resource elements (Resource Element, RE), 1 An RE is one subcarrier within one SC-FDMA symbol time. The basic time-frequency resource structure of the LTE uplink is shown in Figure 1.

LTE uplink control information (Uplink Control Information, UCI) includes HARQ response (ACK/NACK), channel quality indicator (ChannelQuality Indicator, CQI) and scheduling request (Scheduling Request, SR) for downlink data packets, and also includes information for Feedback information related to MIMO, such as rank indicator (Rank Indicator, RI) and precoding matrix indicator (Precoding Matrix Indicator, PMI) of downlink transmission.

UCI is mainly transmitted on the Physical Uplink Control Channel (PUCCH). LTE uses frequency diversity to transmit PUCCH: the first time slot is transmitted in one RB at the edge of the system bandwidth, and the second time slot is transmitted in another RB at the opposite edge of the system bandwidth. The two RBs together are called one The PUCCH domain is shown in Figure 2. According to the type of control information transmitted, PUCCH is divided into seven formats: format 1 only transmits SR; format 1a transmits 1-bit ACK/NACK; format 1b transmits 2-bit ACK/NACK; format 2 only transmits CQI (20 coded bits); Format 2a transmits CQI and 1-bit ACK/NACK (20+1 coded bits); format 2b transmits CQI and 2-bit ACK/NACK (20+2 coded bits).

UCI is also sometimes transmitted on the Physical Uplink Shared Channel (PUSCH). Because a user equipment (User Equipment, UE) cannot simultaneously transmit the PUCCH and the PUSCH in the same subframe. Therefore, when a subframe has been used to transmit PUSCH, the PUCCH cannot be used to transmit control information. At this time, control information such as CQI and ACK/NACK must be multiplexed with data and then transmitted on the PUSCH.

Uplink Control Information (UCI) includes CQI, HARQ response, SR and RI, etc. The HARQ response is obtained from the upper layer and contains 1 or 2 bits. Each positive acknowledgment (ACK) is encoded as a bit "1", and each negative acknowledgment is encoded as a bit "0"; SR and RI are also encoded as 1 or 2 bits; CQI needs to be encoded into 20 or 32 codeword bits.

The CQI of UCI has three formats, which are respectively used for wideband report, subband report configured by high layer and subband report selected by UE. The UE needs to detect the CQI in the physical downlink shared channel (Physical Downlink Shared Channel, PDSCH) before reporting to the base station, so the format of the CQI is different in different PDSCH transmission modes. Each CQI format includes multiple fields, and the number of CQI bits is obtained by adding the bit widths of each field.

UCI uses a (20, A) block code for encoding when transmitting on PUCCH, and uses (32, O) block code and Tail-Biting Convolutional Code (Tail-Biting Convolutional Code) respectively according to the number of CQI bits when transmitting in PUSCH , TBCC) for channel coding. It can be seen that according to the UCI transmission channel and the number of bits transmitted, the user equipment needs to choose between three encoding methods, and different encoding schemes also correspond to different decoding schemes, so the base station receiver also needs to select the corresponding The decoding scheme is used for decoding, and the complexity of the system is relatively high.

Contents of the invention

In order to solve the above technical problems, the present invention proposes a unified encoding method and system for LTE uplink control information, the method replaces (20, A) block code, (32, O) block code and TBCC code with a unified encoding method, The original LTE three channel coding is replaced by a two-stage puncturing of the basic TBCC with a code rate of 1/5, which avoids the choice of multiple coding methods at the user end. Therefore, the base station only needs the TBCC decoding scheme and does not need Selection simplifies the encoding/decoding structure of the system, and its performance is close to or better than LTE block codes, reduces decoding complexity, and improves system performance.

The present invention has taken following technical scheme:

A unified encoding method for LTE uplink control information, the method uses a two-stage puncture of a TBCC with a code rate of 1/5 to complete the channel coding of uplink control information at different lengths, specifically comprising the following steps: a first-level puncture step , use the TBCC generator to encode the uplink control information; the number of the generators used is determined as follows: when the TBCC code is encoded, the generators used are three generators of the LTE convolutional code; When encoding, when (20, A) block code is encoded, the number of bits after TBCC code encoding is the smallest integer greater than or equal to 20; when (32, O) block code is encoded, the bit number after TBCC code encoding The number of bits is the smallest integer greater than or equal to 32, and the number of generators is determined according to the value of the number of bits encoded by the TBCC code; the second-level puncture step, when the second-level puncture is performed on the TBCC code, the first-level puncture is retained The number of bits after all the TBCC codes described in the steps are encoded; when the block code is carried out to the second level of puncturing, a part of the coding bits described in the first level of puncturing step needs to be deleted, and the number of bits to be deleted is determined by the following formula: When the number of bits encoded by the TBCC code is greater than 20, the number of bits deleted is the number of bits encoded by the TBCC code minus 20; when the number of bits encoded by the TBCC code is greater than 32, the number of bits deleted is the number of bits encoded by the TBCC code. Subtract 32 from the number of bits.

A unified encoding system for LTE uplink control information, the system uses a two-stage puncturing of TBCC with a code rate of 1/5 to complete the channel coding of uplink control information at different lengths, specifically including: a first-level puncturing module, using The uplink control information is encoded using the TBCC generator; the number of generators used is determined as follows: when the TBCC code is encoded, the generators used are three generators of the LTE convolutional code; When encoding, when (20, A) block code is encoded, the number of bits after TBCC code encoding is the smallest integer greater than or equal to 20; when (32, O) block code is encoded, the bit number after TBCC code encoding The number of bits is a minimum integer greater than or equal to 32, and the number of generators is determined according to the value of the number of bits encoded by the TBCC code; the second-level puncture module is used to retain the first The number of bits encoded by all TBCC codes described in the first-level puncture step; when the second-level puncture is carried out to the block code, a part of the coded bits described in the first-level puncture step needs to be deleted, and the number of bits to be deleted is given by the following formula Determine: when the number of bits encoded by the TBCC code is greater than 20, the number of bits deleted is the number of bits encoded by the TBCC code minus 20; when the number of bits encoded by the TBCC code is greater than 32, the number of bits deleted is encoded by the TBCC code Subtract 32 from the number of bits after that.

Compared with the prior art, the present invention has the following advantages:

1) The three generators in the code rate 1/5 basic TBCC proposed by the present invention are the same as the TBCC of LTE, so the backward compatibility with LTE can be maintained.

2) In terms of performance, the Frame Error Rate (Frame Error Rate, FER) performance of the two-stage punctured TBCC scheme is close to or better than that of LTE block codes, and the decoding complexity is much lower.

3) The proposed two-level punctured TBCC scheme can use a unified method to encode/decode UCI on PUCCH and PUSCH, which avoids the choice between multiple encoding and decoding methods at the user end and base station end, and simplifies The encoding/decoding structure of the system is defined.

Description of drawings

Fig. 1 is the basic time-frequency resource structure of LTE uplink;

Figure 2 is a PUCCH uplink control structure;

Fig. 3 is the R=1/3 convolutional code in the LTE standard;

Figure 4 is a two-stage perforated TBCC scheme;

Fig. 5(a) ~ Fig. 5(g) the FER performance of the two-level punctured TBCC and LTE block code with a code length of 20 under the UCI of 7 to 13 bits;

Figure 6(a) ~ Figure 6(e) FER performance of the two-stage punctured TBCC and LTE block codes with a code length of 32 when the UCI is 7 to 11 bits;

Figure 7 Comparison of decoding complexity between two-stage punctured TBCC and block code with code length 20

Figure 8 Comparison of decoding complexity between two-stage punctured TBCC and block code with code length 32

Fig. 9 is the base sequence of (20, A) block code;

Fig. 10 is the base sequence of (32, O) block code;

Figure 11 is the perforation pattern when the code length is 20 for the two-stage perforated TBCC;

Figure 12 is the perforation pattern when the code length is 32 for two-stage perforated TBCC;

Figure 13 is a normalized QPSK constellation diagram;

Fig. 14 is the performance when the code length is 20 LTE block codes and two-stage punctured TBCC when FER=10-4;

Fig. 15 shows the performance of an LTE block code with a code length of 32 and a two-stage punctured TBCC when FER=10-4.

Detailed ways

In order to make the above objects, features and advantages of the present invention more comprehensible, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

The present invention provides a unified encoding method for LTE uplink control information. Three of the generators of the basic TBCC with a code rate of 1/5 used in the method are the same as the convolutional codes of LTE to ensure compatibility with the original encoding scheme of LTE. , specifically including two-stage perforation steps. In the first level of puncturing, among the basic TBCC generators with a code rate of 1/5, several or all generators are selected for variable-rate TBCC encoding. When the same three generators as the original LTE convolutional code are selected, it is consistent with the original LTE TBCC code; when replacing the original LTE block code, the number of generators is selected according to the specific number of UCI bits, so as to achieve different encoding rates TBCC, select the number of generators to meet the UCI in different situations, the number of bits obtained after TBCC encoding is greater than/equal to 20 (instead of (20, A) block code) or 32 (instead of (32, O) group code) minimum integer, try to make the bits of the second-level erasure codeword the least. In the second stage of puncturing, the codeword bits output after the first stage of puncturing are deleted to obtain the codeword bit length specified in the original LTE. When replacing the original TBCC code of LTE, the number of deleted bits is zero. When replacing the original block code of LTE, in order to obtain a code word with a length of 20 or 32, and delete bits exceeding the required code length, two methods are used in the second level of puncturing : The first method is to uniformly puncture the encoded bits; the second method is to use the circular buffer in LTE TBCC Rate Matching (Rate Matching, RM) to delete the bits in the output stream of the last generator from the end, namely Cycle delete. The generators and puncturing patterns used in the first-level and second-level puncturing are obtained by searching, and the decoding error probability under all possible puncturing patterns is calculated, and the puncturing pattern with the best error performance is selected.

The present invention uses a two-stage puncture of a basic TBCC with a code rate of 1/5 to complete the channel coding of UCI at different lengths, and the three generators of the basic TBCC with a code rate of 1/5 are the same as the convolutional code of LTE , can maintain backward compatibility with LTE: when replacing the original LTE TBCC code, select the three generators of the original LTE convolutional code in the first level of puncturing, and do not need to delete the codeword in the second level; replace the original LTE block code In the first level of puncturing, the number of TBCC generators is selected according to the number of UCI bits, and TBCC encoding is performed. In the second level of puncturing, the redundant bits after the first level of encoding are deleted to obtain a code length of 20 or 32 codeword. The receiver of the TBCC code can be decoded by an efficient (Wrap Around Viterbi Algorithm, WAVA), and the complexity is much lower than the maximum likelihood (Maximum Likelihood, ML) decoding of the block code, so it can be obtained from both performance and complexity The performance is close to or even better than the original encoding scheme of LTE, and the complexity of the system is reduced. Therefore, UCI can be uniformly encoded by the above two-level punctured TBCC scheme under different lengths, and there is no need to choose between (20, A) block code, (32, O) block code and rate 1/3 TBCC. The base station only needs to perform WAVA decoding on the TBCC code, which is much less complex than the original LTE block code ML decoding.

Example:

The method of the present invention will be described in detail below in conjunction with the accompanying drawings.

1. LTE original encoding method:

When the UCI is transmitted on the PUCCH, a (20, A) block code is used for encoding. The code word of the (20, A) block code of LTE is a linear combination of 13 base sequences, and these 13 base sequences are recorded as (M _{i, 0} , M _{i, 1} ,..., M _{i, 12} ), such as Figure 9 shows. Codeword bits before encoding are denoted as a ₀ , a ₁ , a ₂ , ..., a _A-1 , coded bits are denoted as b ₀ , b ₁ , b ₂ ..., b _B-1 , where B=20 is the number of codeword bits, and each codeword bit is calculated by the following formula:

$b_i=\underset{\mathrm{no}=0}{\overset{A-1}{\mathrm{\&Sigma;}}}(a_{\mathrm{no}}\&CenterDot;m_{i,\mathrm{no}})\mathrm{<figure-callout\; id="2"\; label="mod"\; filenames="HDA0000089623750000011.png,HDA0000089623750000012.png"\; state="\{\{state\}\}">mod</figure-callout>}2,i=\mathrm{0,1,2},...B-1$ [Formula 1]

When the CQI is transmitted on the PUSCH, if it is less than 11 bits, it needs to be encoded into 32 codeword bits. Encoded by a (32, O) variable code rate block code, the CQI bits input to the channel coding module are denoted as o ₀ , o ₁ , o ₂ ,..., o _O-1 , where O is the number of bits . The code word of the (32, O) block code is a linear combination of 11 base sequences, and these 11 base sequences are recorded as (M _{i, 0} , M _{i, 1} ,..., M _{i, 10} ), as shown in Figure 10 shown. The encoded bits are denoted as b ₀ , b ₁ , b ₂ , ..., b _B-1 , where B=32 is the number of codeword bits, and each codeword bit is calculated by the following formula:

$b_i=\underset{\mathrm{no}=0}{\overset{o-1}{\mathrm{\&Sigma;}}}(o_{\mathrm{no}}\&Center\; Dot;m_{i,\mathrm{no}})\mathrm{<figure-callout\; id="2"\; label="mod"\; filenames="HDA0000089623750000011.png,HDA0000089623750000012.png"\; state="\{\{state\}\}">mod</figure-callout>}2,i=\mathrm{0,1,2},...B-1$ [Formula 2]

When the CQI is transmitted on the PUSCH, if it is greater than 11 bits, a TBCC with a constraint length of 7 and a code rate of 1/3 is used for channel coding. The generator of TBCC is (133, 171, 165), and the encoder structure is shown in Figure 3. In order to make the initial state of the shift register the same as the final state, the initial value of the shift register is set to the last 6 information bits of the input information stream. If the input information flow is u ₀ , u ₁ , u ₂ ,..., u _k-1 , k is the number of information bits, and the six storage units in the shift register use s ₀ , s ₁ , s ₂ , s ₃ , s ₄ , s ₅ said, then the initial value of the TBCC shift register should be set to:

s _i =u _k-1-i [Formula 3]

i＝0，1，2。其中D为每个编码输出流的比特数，i表示编码输出流的序号。After encoding, three codeword bit streams are obtained

i=0,1,2. Among them, D is the number of bits of each coded output stream, and i represents the serial number of the coded output stream.

2. According to the original encoding method of LTE, a TBCC generator with a code rate of 1/5 is obtained

In order to maintain the compatibility of LTE, three of the generators of the TBCC with a code rate of 1/5 are the same as the LTE convolutional code, namely (133, 171, 165) ₈ (representing octal form), while the other two generators is obtained by exhaustive search. First, add a generator on the basis of (133, 171, 165) ₈ to form a convolutional code with a code rate of 1/4. For a generator with a constraint length of 7, there are several possibilities as follows:

(1) There is 1 zero in the generator, and there are One case, that is (0111111) ₂ ~ (1111110) ₂ .

种情况，即(0011111)₂～(1111100)₂。(2) There are 2 zeros in the generator for a total of

In this case, (0011111) ₂ ~ (1111100) ₂ .

种情况，即(0001111)₂～(1111000)₂。(3) There are 3 zeros in the generator, a total of

One case, namely (0001111) ₂ ~ (1111000) ₂ .

种情况，即(0000111)₂～(1110000)₂。(4) There are 4 zeros in the generator, a total of

One case, namely (0000111) ₂ ~ (1110000) ₂ .

种情况，即(0000011)₂～(1100000)₂。(5) There are 5 zeros in the generator, a total of

In this case, (0000011) ₂ ~ (1100000) ₂ .

种情况，即(0000001)₂～(1000000)₂。(6) There are 6 zeros in the generator, a total of

One case, namely (0000001) ₂ ~ (1000000) ₂ .

种可能的生成器分别与(133，171，165)_s构成码率为1/4的卷积码，计算出码重分布W(C)：will this

The possible generators and (133, 171, 165) _s form a convolutional code with a code rate of 1/4, and the code redistribution W(C) is calculated:

W(C)=(A _i , i=0, 1, . . . n) [Formula 4]

Wherein, A _i represents the number of all codewords with Hamming weight i in the code C, and n is the code length, where n=32. According to the code weight distribution, the decoding error probability under the Additive White Gaussian Noise (AWGN) channel is calculated as follows:

$P_e\left(C\right)<\underset{w=0}{\overset{\mathrm{no}}{\mathrm{\&Sigma;}}}{A}_{w}Q\left(\sqrt{2\mathrm{wxya}\frac{{\mathrm{E.}}_b}{N_0}}\right)$ [Formula 5]

Among them, R is the code rate, E _b /N ₀ is the signal-to-noise ratio per bit, and Q(x) is the Gaussian Q function. Through searching, the generators with the best decoding error performance for convolutional codes with code rate 1/4 are (055) _s , (132) _s , (113) _s , (151) _s . The generator of the rate 1/4 convolutional code is selected as (133, 171, 165, 132) _s . According to the same method, the generator of the determined rate 1/5 convolutional code is (133, 171, 165, 132 , 157) _s .

3. Based on the basic TBCC generator, implement a two-stage perforated TBCC

The proposed two-stage perforated TBCC scheme is shown in Fig. 4.

1) In the first level of puncturing, UCI of length 7 to 13 bits is encoded using all or part of the basic generator. Assuming that the number of coded bits is K (7≤K≤13), the UCI bits input to the channel coding module are: k ₀ , k ₁ , k ₂ ,..., k _K-1 , the number of basic generators used is N (2≤N≤5), so the encoding rate of TBCC code is 1/N, the number of encoded bits is K*N, and the encoded codeword bits are: b ₀ , b ₁ , b ₂ , .. ., b _K*N-1 , when replacing the original TBCC code of LTE, the three generators N=3 of the LTE convolutional code are selected, and when replacing the LTE block code, the number of K*N is required to be greater than/equal to 20 (in place of (20, A) block code) or 32 (in place of (32, O) block code), thereby determining the number of N.

2) In the second level of puncturing, no deletion is required when replacing the LTE TBCC code, and when replacing the LTE block code, in order to obtain a code word with a code length of 20 or 32, a part of the coded bits (or none) is further deleted, which can be obtained The number of bits P that should be deleted: K*N-20 (K*N>20) or K*N-32 (K*N>32); Two methods have been adopted in the second-level puncture: the first method is to The coded bits are uniformly punctured; the second method is to use the circular buffer in the LTE TBCC rate matching (Rate Matching, RM) to delete the bits in the output stream of the last generator from the end, that is, circular puncture.

4. According to the two-level puncture method in 3, determine the generator and puncture pattern under different UCI bit conditions

The generators and puncture patterns used in the first and second punctures are obtained by searching, and the decoding error probability under all possible puncture patterns is calculated by using formula (5), and the puncture with the best error performance is selected pattern. The final perforation pattern is in Figure 11 and Figure 12.

5. Comparison of the complexity of the decoding method of the two-stage punctured TBCC code and the LTE block code

In summary, the encoding process of the two-stage punctured TBCC scheme is shown in Figure 4. The number of UCI bits is known, and the two-stage puncture is performed according to the parameters given in Figure 11 or Figure 12 to obtain a codeword with a code length of 20 or 32 (Because when replacing the TBCC code, the decoding method is the same, and the decoding complexity is the same, so it will not be discussed here). The receiver can use efficient WAVA for decoding. The comparison of ML decoding and WAVA decoding complexity of LTE block codes is discussed below.

Assume that the two-stage punctured TBCC scheme adopts WAVA for decoding, and the LTE (20, A) block code and (32, O) block code use ML for decoding. WAVA uses the lattice graph of the convolutional code to search for the codeword most similar to the received signal, and ML decoding obtains the codeword most similar to the received signal by comparing the received signal with all possible codewords. When the (n, k) parameters are the same, the complexity of WAVA is much lower than ML decoding in most cases.

Next, compare the complexity of WAVA and ML decoding from the calculation amount of addition operation, multiplication operation and comparison operation. "Addition" refers to the addition of two real values. "Multiplication operation" refers to multiplying the soft information output by the demodulator with "+1" or "-1", which can also be equivalent to checking whether a certain codeword bit is "0" or "1". "Comparison operation" refers to comparing the size of two values and selecting the larger (or smaller) value.

Consider first the WAVA algorithm. Assuming that the maximum number of iterations set during decoding is 2, Viterbi decoding is performed once per iteration, and generally only one iteration is required under high Signal to Noise Ratio (SNR) conditions. The Viterbi algorithm uses the trellis diagram of the convolutional code for decoding: assuming that the length of the information bit is recorded as k, the time range in the trellis diagram is t=1, 2,...,k; the constraint length is recorded as K, then There are 2 ^K states at each moment; the code rate is recorded as 1/q, so each state has 2 ¹ =2 input branches, and there are q codeword bits on each branch. At each moment of Viterbi decoding: first calculate 2 ^q possible branch measures, that is, the distances between 2 ^q possible codewords on the branch and the received signal, and the calculation of each branch measure requires q times of multiplication, q-1 times addition, so there are 2 ^q × q multiplications and 2 ^q × (q-1) additions in total; secondly, for each state, 2 candidate measures are calculated, and the value with the smaller measure is selected as the new state measure, so there are 2 ^K × 2 additions, 2 ^K comparisons. There are k moments in each Viterbi decoding, so the calculation amount of WAVA in one iteration is [2 ^K ×2+2 ^q ×(q-1)]×k additions, 2 ^q ×q×k multiplications and 2 ^K × k comparisons.

Next, consider the ML decoding algorithm. Assuming that the length of the received signal is n, calculating the distance between the received signal and a codeword requires n multiplications and n-1 additions. ML decoding needs to calculate the distance between the received signal and 2 ^k possible code words, and select the code word with the smallest distance, so there are 2 ^k × (n-1) additions, 2 ^k × n multiplications and 2 ^k - 1 comparison.

The above is a coding and decoding method of a unified coding scheme for LTE uplink control information.

In order to evaluate the performance of a unified encoding method for LTE uplink control information of the present invention, the performance of the proposed two-stage punctured TBCC scheme and the LTE original block code scheme is simulated, wherein the TBCC second-level punctured scheme adopts uniform puncture and The loop puncture method is finally compared in terms of performance and complexity:

The encoding process of the two-stage punctured TBCC scheme is shown in Figure 4, and puncture is performed according to the parameters given in Figure 11 or Figure 12 to obtain a codeword with a code length of 20 or 32. The encoding process of the LTE original block coding scheme is shown in formula (1) or formula (2), and the base sequence with a length of 20 or 32 shown in Fig. 9 or Fig. 10 is used respectively. The simulation environment of the channel is an AWGN channel, AWGN noise is generated according to the given SNR, and then the noise is added to the output sequence passing through the encoder and modulator. The modulation method is quadrature phase shift keying (Quadrature PhaseShift Keying, QPSK), using the normalized QPSK constellation shown in Figure 13, and the bit group b(i), b(i+1) is mapped to the complex-valued modulation symbol x=I+jQ. 106 frames of UCI bits were simulated at each SNR. The decoding method of the block code at the receiving end: the two-stage punctured TBCC scheme adopts WAVA for decoding, and the LTE (20, A) block code and (32, O) block code use ML for decoding.

Figure 5(a) to Figure 5(g) show the FER performance of the two-stage punctured TBCC with a code length of 20 and the LTE block code when the UCI is 7 to 13 bits. (n, k) in the figure indicates that the encoder takes k bits as input and outputs n-bit codewords. The performance of each scheme is analyzed with FER=10 ^-4 as the standard, as shown in FIG. 14 . It can be seen from the figure that the TBCC under the cyclic deletion scheme can obtain better performance than the TBCC under the uniform deletion scheme, therefore, the cyclic deletion scheme should be selected for the second-level perforation. When the UCI is 8 and 11 bits, the performance of the TBCC scheme is the same as that of the LTE block code. When the UCI is 10 and 12 bits, the gap between the performance of the TBCC scheme and the block code is very small, only 0.05dB. When the UCI is 7, 9 and 13 bits, the performance of the TBCC scheme is slightly worse than the block code by 0.3dB to 0.5dB.

Figure 6(a) to Figure 6(e) show the FER performance of the two-level punctured TBCC with a code length of 32 and the block code when the UCI is 7 to 11 bits. The performance of each scheme is analyzed with FER=10 ^-4 as the standard, as shown in FIG. 15 . It can be seen from the figure that, as in the case where the code length is 20, the cyclic erasure scheme should be selected for the second level of puncturing to obtain better performance. When the UCI is 7 bits, the performance of the TBCC scheme is 0.35dB better than the LTE block code. When the UCI is 8 bits, the performance of the TBCC scheme is the same as that of the LTE block code. When the UCI is 9 and 11 bits, the gap between the performance of the TBCC scheme and the block code is very small, only 0.05dB or 0.1dB. When the UCI is 10 bits, the performance of the TBCC scheme is slightly worse than the block code by 0.5dB.

Figure 7(a) to Figure 7(c) are the decoding complexity comparisons of the two-stage punctured TBCC with a code length of 20 and the LTE block code under the UCI of 7 to 13 bits, and the TBCC parameters under different UCI lengths are as shown in Figure 11 to set. From Figure 7(a) and Figure 7(b), it can be seen that the complexity of WAVA’s “addition operation” and “multiplication operation” increases linearly with the increase of UCI bits, while the decoding complexity of ML increases with The increase of UCI bits increases exponentially, and the complexity of WAVA is much lower than that of ML. It can be seen from Figure 7(c) that WAVA only needs one iteration under high SNR, and the complexity of "comparison operation" is not much different from ML before 9 bits, but the complexity of WAVA is higher than that of ML starting from 9 bits. Much lower; WAVA requires 2 iterations at low SNR, and the "comparison complexity" is larger than ML before 10 bits, but WAVA has lower complexity than ML after 10 bits.

Figure 8(a) to Figure 8(c) are comparisons of the decoding complexity of the two-stage punctured TBCC with a code length of 32 and LTE block codes with a UCI of 7 to 11 bits, and the TBCC parameters under different UCI lengths are as shown in Figure 12 to set. From Fig. 8(a) and Fig. 8(b), we can see that the complexity of WAVA's "addition operation" and "multiplication operation" is much lower than that of ML. It can be seen from Fig. 8(c) that the "comparative complexity" of WAVA is higher than that of ML before 9 bits and lower than that of ML after 9 bits at 1 iteration, while the "complexity" of WAVA is relatively complex at 2 iterations. Degree" is higher than that of ML and only lower than ML at 11 bits.

The results in Fig. 5, Fig. 6, Fig. 7 and Fig. 8 show that for the FER performance, the proposed two-stage punctured TBCC scheme is not much different from LTE block codes, and even better than LTE block codes at some UCI bit lengths. The two-level punctured TBCC scheme can be decoded by WAVA, and the decoding complexity is much lower than that of block code ML decoding under high SNR. " and "multiplication operations" are still much less complex than ML. Therefore, in terms of both performance and complexity, the performance of the two-stage punctured TBCC scheme is close to or even better than that of the LTE original coding scheme. The proposed two-level punctured TBCC scheme can use a unified method to encode/decode UCI on PUCCH and PUSCH, which avoids the choice between multiple encoding and decoding algorithm schemes at the user end and base station end, and simplifies The compile/decode structure of the system.

The above is a detailed introduction to a unified encoding method and system for LTE uplink control information provided by the present invention. In this paper, specific embodiments are used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only used to help understanding The method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation and application range. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims (2)

1. a unified encoding method of LTE uplink control information, is characterized in that, completes the channel coding of uplink control information under different lengths by two stages of puncturing of the TBCC of a code rate 1/5, specifically comprises the steps: The first level of puncturing step uses the TBCC generator to encode the uplink control information; the number of generators used is determined as follows: when the TBCC code is encoded, the generators used are three generation of LTE convolutional codes device; when block codes are encoded, there are two cases, when (20, A) block codes are encoded, the number of bits after TBCC code encoding is a minimum integer greater than or equal to 20, when (32, O) grouping When the code is encoded, the number of bits encoded by the TBCC code is a minimum integer greater than or equal to 32, and the number of generators is determined according to the value of the number of bits encoded by the TBCC code; The second level of puncturing step, when the TBCC code is carried out to the second level of puncturing, retain the number of bits after all TBCC codes encoded in the first level of puncturing step; when the block code is carried out to the second level of puncturing, the first A part of the coding bits described in the step of perforating needs to be deleted, and the number of bits to be deleted is determined by the following formula: when the number of bits after TBCC code encoding is greater than 20, the number of bits to delete is the number of bits after TBCC code encoding minus 20; When the number of bits encoded by the TBCC code is greater than 32, the number of deleted bits is the number of bits encoded by the TBCC code minus 32. 2. a unified encoding system for LTE uplink control information, characterized in that the system completes the channel coding of uplink control information at different lengths by two stages of puncturing of a TBCC with a code rate of 1/5, specifically comprising: The first-level puncture module is used to encode the uplink control information using the TBCC generator; the number of the generators used is determined as follows: when encoding the TBCC code, the generator used is three of the LTE convolutional code generator; when block codes are encoded, when (20, A) block codes are encoded, the number of bits after TBCC code encoding is a minimum integer greater than or equal to 20; when (32, O) block codes are encoded , the number of bits encoded by the TBCC code is a minimum integer greater than or equal to 32, and the number of generators is determined according to the value of the number of bits encoded by the TBCC code; The second-level puncture module is used to retain the number of bits encoded by all TBCC codes described in the first-level puncture step when the TBCC code is carried out to the second-level puncture; when the block code is carried out to the second-level puncture, A part of the coded bits described in the first-level perforation step needs to be deleted, and the number of bits to be deleted is determined by the following formula: when the number of bits after the TBCC code is encoded is greater than 20, the number of bits to be deleted is the number of bits after the TBCC code is encoded minus 20; when the number of bits encoded by the TBCC code is greater than 32, the number of bits to be deleted is the number of bits encoded by the TBCC code minus 32.

Images