CN101873160A - A method to reduce the amount of feedback when feeding back a sequence composed of multiple codewords - Google Patents

Abstract

All sequences that may be formed by multiple codewords are combined into a set, and when the sequence composed of multiple codewords is fed back, the sequence number of the codeword sequence to be fed back in the set is used for feedback to reduce the amount of feedback when the codeword sequence is fed back .

Description

A method to reduce the amount of feedback when feeding back a sequence composed of multiple codewords

technical field

The technology relates to a method for reducing the amount of feedback data in a codeword feedback system, and in particular to a method for reducing the amount of feedback when feeding back a sequence composed of multiple codewords.

Background technique

In a communication system, the channel state is time-varying, and this time-varying characteristic is reflected in the transfer function of the channel, resulting in the random characteristics of the channel transfer function, that is, the instantaneous effect of the channel on the signal is reflected as a random process. This randomness of the channel will cause the signal to be distorted during transmission, resulting in transmission errors. In modern communication systems, measuring the channel and performing feedback processing on the measurement data are necessary means to realize the prediction of the random characteristics of the channel and use the feedback data to control and adjust the sending and receiving process.

According to actual network deployment and design requirements, this feedback may be transmitted through an uplink control channel or a shared channel. But no matter which way is used, the feedback information to multiple base stations will cause huge uplink overhead. In order to reduce the amount of feedback data, some compromise methods have been proposed: 3GPP’s Release 7 and Release 8 standards in LTE; R1-092530: Required Information at the UE inCoMP; R1-092369: Feedback Design Principles for Downlink CoMP. These methods control the amount of feedback information by reducing the quantization accuracy of the feedback information or the integrity of the feedback information, that is, reduce the amount of feedback data by reducing the accuracy of the feedback information.

In this description, a single user needs to perform information feedback with multiple base stations at the same time. Specifically, the user needs to use the information of each precoding matrix to form a sequence in a certain order when communicating with multiple base stations. All are fed back to the base station.

The base station and the user side store the same precoding codebook respectively. Each precoding matrix corresponds to a codeword in the codebook. The total number of precoding matrices that can be represented in the codebook, that is, the number of codewords included is M. Therefore, when a codeword is represented by a binary number, log ₂ M bits are required to feed back a certain codeword in the codebook. During the feedback process, the user needs to feed back a codeword sequence composed of multiple M codewords to the base station.

The feedback information will be used as a basis for each coordinated base station to schedule users and determine the precoding matrix used for transmission during the user multi-base station cooperation process. Since the precoding matrix needs to match the channel to function, the completeness and accuracy of the feedback information is crucial to the system performance. Using the existing feedback scheme, under the condition of ensuring the accuracy of the feedback information, the user needs to provide complete feedback, which requires a lot of system overhead; while using the existing feedback reduction scheme, since the integrity of the feedback information cannot be guaranteed, This will result in a decrease in system performance. Considering the limitation on the amount of feedback data in actual deployment and the requirement to ensure system performance, a feedback method with low feedback amount that can guarantee the integrity of feedback when feeding back multiple codeword sequences is the most ideal.

Contents of the invention

The invention provides a low-feedback-amount feedback method for feeding back a codeword sequence, which is suitable for a communication system that needs to feed back multiple codewords at the same time. In the present invention, the multiple codewords to be fed back may be the arrangement, combination or sequence of multiple codewords according to specific requirements.

The technical problem to be solved in the present invention is how to use as little feedback data as possible to completely feed back the information contained in the sequence formed by multiple codewords.

In order to solve the above problems, the present invention adopts the following technical solutions: a method for reducing the amount of feedback when feeding back a codeword sequence composed of multiple codewords. The method includes the following steps:

其中

为求从x个元素中取出y个元素进行全排列的个数；针对不需要考虑码字间顺序的情况，码本中的元素为码字的组合。对于由M种码字构成，长度为L的组合，码本中包含的码字组合的个数为

其中

为求从x个元素中取出y个元素进行组合的个数。针对采用特定要求构成的码字序列，不考虑其序列的具体生成步骤，但是需要反馈的发送端与接收端使用相同的序列生成方法。采用该序列生成方法时所有可能产生的码字序列构成集合CB。Step 1: At the sending end and the receiving end of the feedback, the same set generation method is adopted to simultaneously generate a set including all codeword sequences to be fed back as a codebook CB. For the case where the sequence between codewords needs to be considered, the elements in the codebook are the arrangement of codewords. For a permutation consisting of M codewords and a length of L, the number of codeword permutations contained in the codebook is

in

In order to obtain the number of y elements taken out of x elements for full arrangement; for the case where the sequence between codewords does not need to be considered, the elements in the codebook are combinations of codewords. For a combination consisting of M codewords with a length of L, the number of codeword combinations contained in the codebook is

in

In order to find the number of combinations of y elements taken out of x elements. For the code word sequence composed of specific requirements, the specific generation steps of the sequence are not considered, but the sending end and the receiving end that need feedback use the same sequence generation method. All possible codeword sequences that can be generated when this sequence generation method is used constitute a set CB.

Step 2: Traverse the CB with the same element traversal method at the sending end and the receiving end of the feedback, and number each element therein to form a serial number c corresponding to each codeword sequence. In general, sequences and serial numbers can be mapped one-to-one. However, the sequence and the serial number do not necessarily need to satisfy the one-to-one mapping relationship, it only needs to satisfy that each serial number can correspond to a sequence.

Step 3: The sending end of the feedback finds a sequence number c corresponding to the multiple codewords to be fed back in the CB, and uses it as feedback information to feed back to the receiving end of the feedback.

Step 4: According to the received feedback information c, the feedback receiving end finds the codeword sequence corresponding to c in its local codebook CB.

Further, the above method further includes: taking the codeword sequence as the codeword, and iteratively using the solution of the present invention when feeding back the permutation, combination or sequence formed by the codeword sequence according to specific requirements.

Description of drawings

Figure 1 Flowchart.

Detailed ways

The solutions in the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

The core idea of the present invention is to use the set CB composed of all the codeword sequences to store the specific information of the codeword sequences, which are respectively stored at the sending end and the receiving end of the feedback. In the feedback process, only the serial number of the sequence is used as the feedback information, thereby reducing the amount of feedback while ensuring the codeword sequence information.

This embodiment takes precoding feedback under LTE-Advanced CoMP as an example, and describes a technical solution for reducing the amount of feedback when feeding back a sequence composed of multiple codewords.

In the LTE-Advanced CoMP system, the mobile station (UE) needs to feed back to each cooperative cell (Cooperation Point, CP) in the cooperation set (Cooperation Set, CS) the priority arrangement when the cell uses each precoding matrix. Assume that there are 3 CPs in the CS, and assume that the codebook used to feed back the precoding matrix is the precoding matrix codebook containing 16 codewords in the LTE-R8 4-antenna scenario. The UE needs to feed back a sequence containing all 16 codewords to each CP, indicating that the 16 codewords are arranged in descending order of usage priority. At this time, M=16, L=16. If full feedback is performed, the UE needs to feed back L*log ₂ M=64 bits to the base station.

Adopt limited feedback technique among the present invention, adopt following steps:

Step 1. On the CP and UE sides, use (but not limited to) the dictionary order method to generate a full permutation composed of the 16 codewords as a CB.

The algorithm idea of the dictionary order method to generate a full permutation is to specify a sequence relationship for the characters in a given character set. Determine the next permutation of the full permutation, and sequentially generate the next permutation of the full permutation until all the permutations are generated. The so-called next of one is that there is nothing between this one and the next. This requires that this one has the longest possible common prefix with the next one, that is, the change is limited to the shortest possible suffix. Thus, according to this method, for a given character set, all possible full permutations are enumerated without repetition. The generation method of this full arrangement is the dictionary order method.

Step 2. Number each permutation in turn according to the order in which the permutations are generated. In the dictionary order method, the sequence of different sequences is determined by comparing the sequence of contents in each position from left to right.

Step 3. The UE compares the codeword sequences corresponding to the multiple precoding matrices to be fed back with each element in the CB, finds out the number c corresponding to the codeword sequence in the CB, and feeds it back to the serving base station.

Step 4. According to the received number c, the base station searches for the corresponding code sequence in its own CB, and uses it as the received code sequence for feedback.

When binary numbers are used to fully represent the number of each arrangement, each feedback data needs a total of bit.

Under the condition that the amount of feedback information is guaranteed to be consistent with that of complete feedback, the method proposed by the invention can greatly reduce the amount of feedback data.

The specific embodiments to which the present invention is applicable are not limited to the precoding feedback system. It can also be applied to quantization-based channel quality indicator (Channel Quality Indicator, CQI) sequence feedback, modulation and coding mode selection (Modulation and Coding Selection, MCS) information sequence feedback, etc. The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person familiar with the technology can easily think of changes or replacements within the technical scope disclosed in the present invention. , should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims (6)

1. A method for reducing the amount of feedback when feeding back a sequence composed of a plurality of codewords, characterized in that the method package comprises steps: A. Form a set CB of sequences composed of multiple codewords to be fed back, the number of codewords contained in it is [CB], and [X] is the number of elements in the calculation set X. B. Traverse CB, and number each element in it, and obtain the sequence number c corresponding to each codeword sequence C. At the sending end of the feedback, find out the sequence number c corresponding to the sequence formed by the plurality of codewords to be fed back in the CB, and use it as feedback information to feed back; D. At the feedback receiving end, according to the obtained sequence number c, obtain a sequence composed of multiple codewords from the CB. 2. for a kind of method described in claim 1 when feeding back the sequence that is formed by a plurality of codewords, reduce feedback amount, it is characterized in that, step A further comprises following three kinds of situations: A1. For the situation where the sequence between codewords needs to be considered, at the feedback sending end and the receiving end, a permutation method is used to generate a set CB containing all full permutations composed of multiple codewords, and the full permutation of the codewords contained therein of at most

in

In order to find the number of y elements taken out of x elements for full arrangement; A2. For the situation where the sequence between codewords does not need to be considered, at the feedback sending end and the receiving end, a permutation method is adopted to generate a set CB containing all combinations of codewords, and the number of combinations of codewords contained therein at most in

In order to find the number of combinations of y elements taken out of x elements; A3. For the codeword sequence formed according to special requirements, the specific generation steps of the sequence are not considered, but the sending end and the receiving end that need feedback use the same sequence generation method to generate multiple codewords when using this sequence generation step All codeword sequences that may be generated form a set CB. 3. A method for reducing the amount of feedback when feeding back a sequence composed of a plurality of codewords described in claim 1, is characterized in that, the mode of obtaining sequence number in step B has the following characteristics: B1. Both the user and the base station use a consistent method to traverse the elements in the CB set, and obtain the sequence numbers corresponding to each sequence in the set before feeding back. 4. A method for reducing the amount of feedback when feeding back a sequence composed of a plurality of codewords described in claim 3, wherein the correspondence between the sequence and the serial number in the mode C1 satisfies: C11. Each sequence number corresponds to a sequence; C12. Different sequence numbers can correspond to the same sequence. 5. A method for reducing the amount of feedback when feeding back a sequence composed of multiple codewords described in claim 1, characterized in that, when there are multiple codeword sequences that need to be fed back, the method in claim 1 Can be used iteratively. 6. A method for reducing the amount of feedback when feeding back a sequence composed of a plurality of codewords described in claim 5, is characterized in that the specific steps of iterative use are: A. Use each codeword sequence as a new codeword, and form a set CB of all possible sequences formed by each codeword sequence. The sequence formed by the code sequence can be the arrangement, combination of the code sequence, or a sequence formed according to special requirements; B. The sending end and the receiving end of the feedback adopt the same sequence set generation and traversal method; C. The sending end of the feedback finds the sequence number c corresponding to it in the set CB according to the sequence formed by the codeword sequence to be fed back, and feeds back the sequence number; D. The receiving end of the feedback searches the corresponding sequence in the CB according to the received sequence number c, and uses it as the feedback result.