Extensible distributed receiving and simultaneous digital multi-beam closed-loop series processing architecture and method

Abstract

The invention provides an extensible distributed simultaneous receiving digital multi-beam forming closed-loop series processing architecture and method, which are used for realizing distributed digital multi-beam forming of large-scale array low data interaction. The architecture comprises M identical processing sub-modules, and each processing sub-module outputs L beams to form LM simultaneous multi-beams. And each processing sub-module is serially connected with two adjacent processing sub-modules to form a series closed loop structure. And each processing sub-module performs sub-array beam forming of LM beams of local P unit antennas, timestamps data after beam forming, and performs caching processing and beam forming according to design requirements. The beam former adopts a closed-loop series architecture, and the interaction data rate between processing submodules is only related to the sampling data rate S and the number of digital beams formed at the same time. The method is suitable for low-cost and rapid scale expansion of medium and large scale digital array multi-beam antennas.

Description

Technical Field

[0001] This application relates to the technical field of digital beamforming, and particularly to a scalable distributed receiving simultaneous digital multi-beam closed-loop cascade processing architecture and method. Background Art

[0002] The significant increase in the scale of digital array antenna elements, the increase in the total baseband data volume, and the increase in the complexity of digital beamforming processing have made the improvement of system expansion complexity an urgent problem to be solved. Existing large-scale digital arrays generally adopt a multi-beamformer architecture with a multi-stage pipelined structure, that is, it consists of two parts: digital sub-array level processing and centralized processing. As the scale of the array increases, the data aggregation and processing pressure of the centralized processing part increase linearly, which is not conducive to the modular design and scale expansion of large-scale arrays. Therefore, it is particularly necessary to study the processing architecture for the distributed implementation of simultaneous multi-beamforming and its scale expansion of large-scale digital receiving array antennas. Summary of the Invention

[0003] This application provides a scalable distributed receiving simultaneous digital multi-beam closed-loop cascade processing architecture and method, which can be used to solve the technical problem of difficult modularization of large-scale arrays.

[0004] This application provides a scalable distributed simultaneous receiving digital multi-beamforming closed-loop cascade processing architecture, which is composed of M identical processing sub-modules and is used for simultaneous multi-beamforming processing of a receiving digital array antenna with an array scale of N = PM unit antennas. Each processing sub-module is responsible for collecting baseband data of P antenna receiving channels;

[0005] The architecture is serially connected head-to-tail by M sub-modules through a point-to-point data interaction interface to form a closed-loop structure;

[0006] The architecture can simultaneously form K = LM independent digital multi-beams, and each processing sub-module completes sub-array beamforming of K beams for P antenna units;

[0007] Each processing sub-module locally caches the sub-array beamforming data of K = LM beams with time stamps, that is, Time labels, according to preset requirements, and accumulates them respectively with the partial sub-array beamforming data of L(M - 1) beams received from the previous processing sub-module according to the corresponding Time labels;

[0008] Each processing sub-module directly outputs the full-array beamforming data of local L beams to the subsequent processing system, and sends the partial sub-array beamforming data of the remaining L(M - 1) beams to the next processing sub-module.

[0009] Further, the functions of the input and output interfaces of each processing sub-module include:

[0010] The input interface is used for the baseband data after digital sampling of P antenna receiving channels, with a sampling data rate of R and an input data rate of PR; and for receiving the beamformed data of L(M - 1) beams from the previous processing sub-module, with an average input data rate of L(M - 1)R.

[0011] Output interface: It is used to output the fully - array beamformed data of local L beams to the subsequent processing system, with an output data rate of LR; and to send the partially - sub - array beamformed data of L(M - 1) beams to the next processing sub-module, with an output data rate of L(M - 1)R.

[0012] On the other hand, the present application also provides an extensible distributed receiving simultaneous digital multi - beam closed - loop cascade processing method, which is implemented by the architecture of the present application. The processing method of each processing sub-module includes:

[0013] Step 1, allocate ID numbers: Each processing sub-module is respectively allocated an ID number m, where m = 1, 2, …, M.

[0014] Step 2, allocate weight coefficients and beam grouping.

[0015] Step 3, local beamforming: The processing sub-module m receives the baseband data of local P antenna receiving channels through the input interface and performs sub - array - level beamforming of K = LM beams.

[0016] Step 4, data caching: Divide the K = LM beams into M groups, with L beams in each group; to ensure the program consistency of each processing sub-module, each processing sub-module caches the sub - array beamformed data in order by beam group.

[0017] Step 5, beam synthesis and output.

[0018] Further, in the extensible distributed simultaneous receiving digital multi - beamforming closed - loop cascade processing method of the present application, Step 2, allocate weight coefficients and beam grouping, includes:

[0019] Divide the K = LM beams into M groups, with L beams in each group; the allocation of M beam groups of each processing sub-module is implemented by the weight coefficient configuration link; the beam group responsible for output by the processing sub-module m is denoted as Lm; the weight coefficients of the M beam groups of each processing sub-module are configured in a cyclic shift manner by beam group, ensuring that the starting beam group of the sub - array beamforming module of the processing sub-module m is always the beam group Lm; through the weight coefficient configuration link, the weight configuration of the M beam groups of each processing sub-module is performed, so that the beam group Lm in the processing sub-module m is always located at the starting position of the beam groups of this processing sub-module.

[0020] Further, in step 4, the method for caching the data after subarray beamforming is as follows:

[0021] The first beam group caches M - 1 packets of data, that is, delays M - 1 Time labels. The second beam group caches M - 2 packets of data, and so on in decreasing order. The Mth beam group does not perform caching.

[0022] Further, step 5, beam synthesis and output, includes:

[0023] The processing sub - module m waits for the data after subarray - level beamforming with the same Time label in other processing sub - modules, that is, it is passed and accumulated successively from the processing sub - module m + 1 to the next processing sub - module, and finally loop - inputs to the processing sub - module m, thus completing the full - array beamforming of the beam group Lm; the data after the full - array beamforming of the beam group Lm is directly output to the subsequent processing module through the output interface.

[0024] Compared with the prior art, the present invention has the following advantages:

[0025] (1) The present invention provides an extensible distributed receiving simultaneous digital multi - beam closed - loop cascade processing architecture and method, realizing load balancing among processing sub - modules and distributed on - array processing. Each individual processing sub - module only interacts with the two adjacent processing sub - modules before and after. The data interaction rate is only related to the number of beams and the baseband data sampling rate, and has nothing to do with the array scale; (2) For the distributed processing architecture proposed by the present invention, the processing structure of each processing sub - module is exactly the same, with strong extensibility. The system can be expanded by inserting more processing sub - modules, and the software of the sub - modules does not need to be modified.

[0026] The present invention also has some defects. For example, the data processing delay increases with the increase in the number of processing sub - modules. The entire beamforming system has strict requirements on the reliability of each processing sub - module, that is, when a certain intermediate node fails, the entire system will be affected. Description of the Drawings

[0027] Figure 1 It is a diagram of an extensible distributed simultaneous receiving digital multi - beam forming closed - loop cascade processing architecture;

[0028] Figure 2 It is a diagram of the internal processing method of the processing sub - module in the extensible distributed receiving simultaneous digital multi - beam closed - loop cascade processing architecture;

[0029] Figure 3 It is a schematic diagram of a specific example of the closed - loop cascade processing architecture under the conditions of N = 32, M = 4, K = 16, and R = 64 MSPS;

[0030] Figure 4Schematic diagram of an example of the internal processing method of a processing sub-module in the case of N = 32, M = 4, K = 16, and R = 64 MSPS. Detailed implementation manners

[0031] To make the objectives, technical solutions, and advantages of the present application clearer, the following will further describe the embodiments of the present application in detail with reference to the accompanying drawings.

[0032] The present application uses a series of closed-loop cascaded processing sub-modules to implement sub-array distributed on-array processing for simultaneous digital multi-beamforming in reception, such that the data interaction data rate between the processing sub-modules is independent of the array element scale and is only related to the number of beams and the signal bandwidth. It is particularly suitable for the implementation and scale expansion of large-scale digital receiving array antennas.

[0033] The following will describe the present invention in detail with reference to the accompanying drawings.

[0034] For a digital receiving array with the number of array elements N = PM, where P is the number of unit antennas corresponding to each processing sub-module and M is the total number of processing sub-modules, and the number of receiving beams K = LM, where L is the number of beams in each group. The beamforming output signal vector is expressed as:

[0035] y(t) = W H x(t) (1)

[0036] Wherein,

[0037] y(t) = [y1(t) y2(t) … y k (t) … y K (t)] T (2)

[0038] x(t) = [x1(t) x2(t) … x n (t) … x N (t)] T (3)

[0039] x(t) is an N×1 digital array element input vector,

[0040] W = [w1 w2 … w k … w K (4)

[0041] W is an N×K multi-beam weight coefficient matrix,

[0042] w k = [w k,1 w k,2 … w k,n … w k,N T (5) ​

[0043] is the weight coefficient vector for the k-th beam.

[0044] First, perform matrix partitioning in the element dimension. Divide all elements into M blocks, and each block corresponds to a sub-array with P = N / M elements. The element data after partitioning is expressed as:

[0045]

[0046] where the input of the m-th (m = 1, 2, 3, …, M) sub-array elements is:

[0047] x m (t) = [x P(m-1)+1 (t) x P(m-1)+2 (t) … x Pm (t))] T (7)

[0048] The weight coefficient matrix of the K beams corresponding to the m-th sub-array is:

[0049] W m = [w 1,k w 2,m … w k,m … w K,m (8)

[0050] where

[0051] w k,m = [w k,P(m-1)+1 w k,P(m-1)+2 … w k,Pm T (9)

[0052] The output signal vector of the K beams of the m-th sub-array is:

[0053]

[0054] The beam output y(t) of the entire array is expressed as:

[0055]

[0056] Through the block processing in the element dimension, the computational pressure of baseband data reception and digital beamforming is evenly distributed among the M sub-arrays. Each processing sub-module is responsible for collecting the baseband data after digital sampling of P antenna receiving channels. The baseband data sampling rate is R, and the input data rate is PR. Each processing sub-module is responsible for sub-array level beamforming of P elements for all K beams.

[0057] ​Subsequently, perform block processing from the beam dimension, divide the K = LM beams into M groups, and each processing sub-module is responsible for the data output after full matrix beamforming of L beams. Assign an ID number to each processing sub-module, where m = 1, 2, …, M. The beam group assigned to processing sub-module m is denoted as Lm. All processing sub-modules form a closed-loop series architecture through a serial connection method, as shown in Figure 1 shown. For beam group Lm, according to Equation (11), starting from processing sub-module m + 1, along the closed-loop structure, the data after sub-array level beamforming of this beam group is sequentially passed downwards, and is accumulated with the data after sub-array level beamforming of the beam group Lm of the subsequent processing sub-module until the closed-loop returns to processing sub-module m. Processing sub-module m accumulates the data after partial sub-array beamforming of beam group Lm received with the data after local sub-array beamforming, completes the beamforming of all sub-arrays, and outputs it to the subsequent processing system through the output interface.

[0058] The data flow direction of the data after sub-array beamforming of the beam group Lm assigned to each processing sub-module starts from the next processing sub-module, flows back to this processing sub-module along the closed-loop structure, and is output to the subsequent processing system. As shown in Figure 2 shown.

[0059] In the closed-loop series architecture, the data rate of the data after beamforming that each processing sub-module needs to receive is (M - 1) * L * R, and the data rate of the data after beamforming output is LM * R, where the data rate of the data after full matrix beamforming output to the subsequent processing system is L * R, and the data rate of the data after sub-array beamforming output to the next processing sub-module is (M - 1) * L * R. The data rate of the data after beamforming input and output by each processing sub-module is only related to the number of beams K and the sampling data rate R, and has nothing to do with the array scale. As the array scale increases, more processing sub-modules can be added to the closed-loop series architecture to expand the system to achieve simultaneous multi-beamforming processing of a large-scale digital receiving array.

[0060] For the above-expandable distributed simultaneous receiving digital multi-beamforming closed-loop series processing architecture, each processing sub-module is serially connected through a point-to-point data interaction interface for distributed on-array processing. The difficulty in its implementation lies in ensuring data synchronization among processing sub-modules. The present invention uses the method of inserting a Time label into the data after beamforming for data alignment, and the specific method is as follows:

[0061] (1) Processing sub-module m receives the baseband data of the local P antenna receiving channels through the input interface, performs sub-array level beamforming of K = LM beams, and attaches a Time label according to the data packet.

[0062] (2) Each processing sub-module locally caches the sub-array beamforming data of K beams with Time labels according to the design requirements. Specifically, the data after sub-array level beamforming of beam group Lm in processing sub-module m needs to cache M - 1 packets of data, that is, delay M - 1 Time labels, for waiting for the data after sub-array level beamforming of beam group Lm to be transmitted and accumulated sequentially to the next processing sub-module starting from processing sub-module m + 1 along the closed-loop series structure, and finally input to processing sub-module m in a closed loop to complete the full-array beamforming of beam group Lm.

[0063] (3) The remaining M - 1 beam groups are cached according to the number of processing sub-modules between their respective corresponding processing sub-modules and processing sub-module m. Specifically, beam group Lm - 1 does not need to cache, beam group Lm - 2 needs to cache 1 packet of data, and so on, beam group L1 needs to cache m - 2 packets of data; due to the series closed-loop structure, in front of processing sub-module 1 is processing sub-module M, beam group LM needs to cache m - 1 packets of data,..., beam group Lm + 1 needs to cache M - 2 packets of data. The data after sub-array level beamforming of the remaining M - 1 beam groups is accumulated corresponding to the data after sub-array level beamforming of the previous processing sub-module m - 1 received, and is output to the next processing sub-module m + 1 through the output interface.

[0064] For the research and development requirements of an actual simultaneous digital multi-beam smart antenna system, taking a system with the number of antenna elements N = 32, the number of processing sub-modules M = 4, the number of receiving beams K = 16, and the baseband data sampling rate R = 64 MSPS as an example, a specific case of a scalable distributed simultaneous digital multi-beamforming closed-loop series processing architecture is given, and the processing framework is as Figure 3 shown. The 16 beams are grouped according to the number of processing sub-modules, with L = 4 beams in each group, and each processing sub-module is responsible for the output of the data after its full-array beamforming.

[0065] Each processing sub-module obtains the baseband data of local P = 8 receiving antenna elements through the input interface, performs sub-array beamforming processing for K = 16 beams and stamps the time on the data after beamforming. Each processing sub-module locally caches the sub-array beamforming data of K = 16 beams with Time labels according to the design requirements. Taking processing sub-module 1 as an example, the data after sub-array level beamforming of 4 beams in its beam group L1 needs to cache 3 packets of data, that is, delay 3 Time labels, for waiting for the data after sub-array level beamforming of part of beam group L1 that is transmitted and accumulated sequentially to the next processing sub-module starting from processing sub-module 2 along the closed-loop series structure and finally input to processing sub-module 1 in a closed loop, complete the accumulation, obtain the data after full-array beamforming of beam group L1, and output it to the subsequent processing system.

[0066] The remaining three beam groups on processing sub-module 1 perform data caching and processing according to their respective closed-loop positions. That is to say, beam groups L2 and L3 cache 2 packets of data and 1 packet of data respectively, and after being accumulated with the data of the corresponding beam group received by processing sub-module 4, they are output to processing sub-module 2. The last L4 beam group does not need to be cached as it is the start of beamforming and is directly output to processing sub-module 2, as Figure 4 shown.

[0067] In the closed-loop cascade architecture, the data rate of each processing sub-module receiving the baseband data of P antenna channels is P*R = 512 MSPS, and the average data rate of the beamformed data received from the previous processing sub-module is (M - 1)*L*R = 768 MSPS; the average data rate of the full-array beamformed data output to the subsequent processing system is L*R = 256 MSPS, and the average data rate of the sub-array beamformed data output to the next processing sub-module is (M - 1)*L*R = 768 MSPS.

[0068] The present invention provides a scalable distributed simultaneous digital multi-beam processing closed-loop cascade architecture, which realizes the load balancing of data transmission among processing sub-modules and distributed on-array processing. The data rate of interaction between each processing sub-module is only related to the number of beams and the baseband data sampling rate, and has nothing to do with the array scale. In the context of large-scale array applications, it solves the bottlenecks of data transmission bandwidth and centralized processing existing in the centralized processing architecture; and the present invention has the characteristic of scalability, and can adapt to arrays of different scales by flexibly adjusting the number of sub-modules in the architecture, which is very suitable for establishing a general simultaneous digital multi-beam processing platform

[0069] The above-described embodiments of the present application do not constitute a limitation on the protection scope of the present application.

Claims

1. An extensible distributed closed-loop tandem processing architecture for simultaneous reception digital multi-beamforming, characterized in that: The architecture consists of M identical processing sub-modules, which are used for simultaneous multi-beamforming processing of a receiving digital array antenna with an array scale of N = PM unit antennas. Each processing sub-module is responsible for collecting baseband data of P antenna receiving channels; The architecture forms a closed-loop structure by serially connecting the M sub-modules head to tail through a point-to-point data interaction interface; The architecture can simultaneously form K = LM independent digital multi-beams. Each processing sub-module completes sub-array beamforming of K beams for P antenna elements; Each processing sub-module locally caches the sub-array beamforming data of K = LM beams with time stamps (i.e., Time labels) according to preset requirements, and accumulates them respectively with the partially sub-array beamforming data of L(M - 1) beams received from the previous processing sub-module according to the corresponding Time labels; Each processing sub-module directly outputs the full-array beamforming data of the local L beams to the subsequent processing system, and sends the partially sub-array beamforming data of the remaining L(M - 1) beams to the next processing sub-module.

2. The scalable distributed simultaneous reception digital multi-beamforming closed-loop cascade processing architecture according to claim 1, wherein The functions of the input and output interfaces of each processing sub-module include: The input interface is used for the baseband data after digital sampling of P antenna receiving channels, with a sampling data rate of R and an input data rate of PR; and for receiving the beamforming data of L(M - 1) beams from the previous processing sub-module, with an average input data rate of L(M - 1)R; The output interface: is used for outputting the full-array beamforming data of the local L beams to the subsequent processing system, with an output data rate of LR; and for sending the partially sub-array beamforming data of L(M - 1) beams to the next processing sub-module, with an output data rate of L(M - 1)R.

3. A scalable distributed receiving simultaneous digital multi-beam closed-loop series processing method, which is implemented using any one of the architectures in claims 1 to 2. The processing method of each processing sub-module includes: Step 1, assign ID numbers: Assign ID numbers m to each processing sub-module, where m = 1, 2,..., M; Step 2, assign weight coefficients and beam grouping; Step 3, local beamforming: The processing sub-module m receives the baseband data of the local P antenna receiving channels through the input interface and performs sub-array level beamforming of K = LM beams; Step 4, data caching: Divide the K = LM beams into M groups, with L beams in each group; Each processing sub-module caches the sub-array beamforming data in order in units of beam groups; Step 5, beam synthesis and output.

4. The scalable distributed simultaneous reception digital multi-beamforming closed-loop series processing method according to claim 3, wherein Step 2, assign weight coefficients and beam grouping, including: Divide K = LM beams into M groups, with L beams in each group; the allocation of the M beam groups of each processing sub-module is realized by the weight coefficient configuration link; the beam group responsible for output by processing sub-module m is denoted as Lm; the weight coefficients of the M beam groups of each processing sub-module are configured in a cyclic shift manner in units of beam groups, ensuring that the starting beam group of the sub-array beamforming module of processing sub-module m is always beam group Lm; the weight configuration of the M beam groups of each processing sub-module is performed through the weight coefficient configuration link, so that beam group Lm in processing sub-module m is always located at the starting position of the beam groups of this processing sub-module.

5. The scalable distributed simultaneous reception digital multi-beamforming closed-loop cascade processing method according to claim 3, characterized in that, In step 4, the method for caching the data after sub-array beamforming is as follows: The first beam group caches M - 1 packets of data, that is, delays M - 1 Time labels. The second beam group caches M - 2 packets of data, decreasing in sequence. The Mth beam group does not perform caching.

6. The scalable distributed simultaneous reception digital multi-beamforming closed-loop cascaded processing method according to claim 3, characterized in that Step 5, beam synthesis and output, includes: Processing sub-module m waits for the data after sub-array level beamforming with the same Time label in other processing sub-modules, that is, it is transmitted and accumulated to the next processing sub-module in sequence starting from processing sub-module m + 1, and finally looped and input to processing sub-module m, thereby completing the full-array beamforming of beam group Lm; the data after the full-array beamforming of beam group Lm is directly output to the subsequent processing module through the output interface.

Images