CN116418381B - Parallel digital multi-beam synthesis method and device based on GPU computing platform - Google Patents
Parallel digital multi-beam synthesis method and device based on GPU computing platform Download PDFInfo
- Publication number
- CN116418381B CN116418381B CN202310226191.2A CN202310226191A CN116418381B CN 116418381 B CN116418381 B CN 116418381B CN 202310226191 A CN202310226191 A CN 202310226191A CN 116418381 B CN116418381 B CN 116418381B
- Authority
- CN
- China
- Prior art keywords
- gpu
- digital multi
- antenna data
- target
- time unit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001308 synthesis method Methods 0.000 title claims abstract description 25
- 230000015572 biosynthetic process Effects 0.000 claims abstract description 97
- 238000003786 synthesis reaction Methods 0.000 claims abstract description 97
- 238000012545 processing Methods 0.000 claims abstract description 29
- 238000000034 method Methods 0.000 claims abstract description 22
- 230000010354 integration Effects 0.000 claims description 44
- 238000001228 spectrum Methods 0.000 claims description 38
- 238000009825 accumulation Methods 0.000 claims description 9
- 230000005540 biological transmission Effects 0.000 claims description 9
- 230000002194 synthesizing effect Effects 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 5
- 238000000638 solvent extraction Methods 0.000 claims 2
- 238000004364 calculation method Methods 0.000 abstract description 13
- 230000001133 acceleration Effects 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 10
- 238000013461 design Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- HPTJABJPZMULFH-UHFFFAOYSA-N 12-[(Cyclohexylcarbamoyl)amino]dodecanoic acid Chemical compound OC(=O)CCCCCCCCCCCNC(=O)NC1CCCCC1 HPTJABJPZMULFH-UHFFFAOYSA-N 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/06—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
- H04B7/0613—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission
- H04B7/0615—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal
- H04B7/0617—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal for beam forming
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/08—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station
- H04B7/0837—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station using pre-detection combining
- H04B7/0842—Weighted combining
- H04B7/086—Weighted combining using weights depending on external parameters, e.g. direction of arrival [DOA], predetermined weights or beamforming
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Variable-Direction Aerials And Aerial Arrays (AREA)
Abstract
The invention provides a parallel digital multi-beam synthesis method and a device based on a GPU computing platform, wherein the method comprises the following steps: transmitting a plurality of groups of beam coefficients and multi-path antenna data from a Central Processing Unit (CPU) to a Graphic Processing Unit (GPU); in the GPU, digital multi-beam synthesis is performed based on the plurality of groups of beam coefficients and the multi-path antenna data; and transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage. The data for digital multi-beam synthesis is transmitted to the GPU for parallel acceleration processing, and the obtained digital multi-beam synthesis result is transmitted back to the CPU for storage, so that the beam synthesis time can be remarkably reduced, the calculation efficiency under large-scale data is improved, and the real-time performance of astronomical observation is ensured.
Description
Technical Field
The invention relates to the technical field of beam synthesis, in particular to a parallel digital multi-beam synthesis method and device based on a GPU computing platform.
Background
In large radio array digital signal processing systems, digital beam synthesis is a key technology, mainly used for array data processing. The digital multi-beam synthesis is to acquire the spatial information of antennas with different paths at each position, weight the signals of each antenna by using a weighting factor, and strengthen or restrain the signals. By utilizing the technology, a plurality of main beams can be formed, and simultaneous observation of a plurality of sky power sources is realized.
With the development of science and technology, the observation equipment in the astronomical field is more advanced, the observation data is more and more, the bandwidth of more than ten Gb can be achieved by the data flow collected in real time, and the problems of long digital multi-beam synthesis calculation time, low efficiency and the like under large-scale data limit the instantaneity and the efficiency of astronomical observation.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a parallel digital multi-beam synthesis method and device based on a GPU computing platform.
In a first aspect, the present invention provides a parallel digital multi-beam synthesis method based on a graphics processor GPU computing platform, comprising:
transmitting a plurality of groups of beam coefficients and multi-path antenna data from a Central Processing Unit (CPU) to a Graphic Processing Unit (GPU);
in the GPU, digital multi-beam synthesis is performed based on the plurality of groups of beam coefficients and the multi-path antenna data;
and transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage.
Optionally, the performing digital multi-beam synthesis based on the multiple sets of beam coefficients and the multiple antenna data includes:
for synthesis of any target beam in any target integration time period, acquiring a set of beam coefficients for synthesizing the target beam and a set of multi-path antenna data in each time unit contained in the target integration time period based on the plurality of sets of beam coefficients and the multi-path antenna data;
respectively carrying out weighted summation on a group of multipath antenna data in each time unit based on the group of beam coefficients;
obtaining a power spectrum of the target beam in a corresponding time unit based on the weighted summation result respectively corresponding to each time unit and the multipath antenna data autocorrelation result summation item respectively corresponding to each time unit;
and accumulating the power spectrums of the target beam in each time unit to obtain a synthesis result of the target beam.
Optionally, the obtaining the power spectrum of the target beam in the corresponding time unit based on the weighted summation result respectively corresponding to each time unit and the multiple paths of antenna data autocorrelation result summation items respectively corresponding to each time unit includes:
determining a power spectrum of the target beam within any target time unit based on the following formula:
where A represents the power spectrum of the target beam over the target time unit,representing the weighted sum result corresponding to said target time unit,/->And representing a multi-channel antenna data autocorrelation result accumulation item corresponding to the target time unit.
Optionally, the GPU includes at least one kernel, and each kernel is configured to perform digital multi-beam synthesis during an integration period based on multiple antenna data and multiple sets of beam coefficients during the integration period.
Optionally, each kernel includes a plurality of thread blocks, and each block includes a plurality of thread;
wherein each said thread is used to calculate the power spectrum of a beam in a time unit.
Optionally, the block division manner in each kernel is determined based on the number of time units and the number of frequency points in one integration period, and the thread division manner in each block is determined based on the number of time units and the number of beam forming of one frequency point in one integration period.
Optionally, the method further comprises:
and the step of transmitting the plurality of groups of beam coefficients and the plurality of paths of antenna data from the CPU to the GPU, the step of performing digital multi-beam synthesis and the step of transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage are taken as three stages of a pipeline, and the digital multi-beam synthesis is realized by adopting a pipeline parallel processing mode.
In a second aspect, the present invention also provides a parallel digital multi-beam synthesis apparatus based on a GPU computing platform, including:
the first transmission module is used for transmitting a plurality of groups of beam coefficients and multiple paths of antenna data from a Central Processing Unit (CPU) to the GPU;
the beam synthesis module is used for carrying out digital multi-beam synthesis in the GPU based on the plurality of groups of beam coefficients and the multi-path antenna data;
and the second transmission module is used for transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage.
In a third aspect, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the parallel digital multi-beam synthesis method based on the GPU computing platform according to the first aspect, when the program is executed by the processor.
In a fourth aspect, the present invention also provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor implements the parallel digital multi-beam synthesis method based on a GPU computing platform as described in the first aspect.
According to the parallel digital multi-beam synthesis method and device based on the GPU computing platform, data for digital multi-beam synthesis are transmitted to the GPU to be subjected to parallel acceleration processing, and the obtained digital multi-beam synthesis result is transmitted back to the CPU to be stored, so that the beam synthesis time can be remarkably reduced, the computing efficiency under large-scale data is improved, and the real-time performance of astronomical observation is ensured.
Drawings
In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a parallel digital multi-beam synthesis method based on a GPU computing platform provided by the invention;
fig. 2 is a schematic diagram of an implementation of a parallel digital multi-beam synthesis method based on a GPU computing platform according to the present invention;
fig. 3 is a schematic diagram of a design of parallel digital multi-beam synthesis of antenna data in a kernel according to the present invention;
FIG. 4 is a schematic diagram of a data processing pipeline design provided by the present invention;
fig. 5 is a schematic structural diagram of a parallel digital multi-beam synthesizing device based on a GPU computing platform according to the present invention;
fig. 6 is a schematic structural diagram of an electronic device provided by the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Graphics processors (Graphics Processing Unit, GPUs) have become the first acceleration device in the field of high performance parallel computing, where an important means of resolving parallel computing using GPUs is to use a compute unified device architecture (Compute Unified Device Architecture, CUDA) architecture.
Aiming at the problems of long calculation time and low efficiency of digital multi-beam synthesis under the real-time observation of a large radio array, the invention provides a parallel digital multi-beam synthesis scheme based on a central processing unit (Central Processing Unit, CPU) +GPU heterogeneous calculation platform, so as to reduce the beam synthesis time, improve the calculation efficiency under large-scale data and ensure the real-time of astronomical observation.
Fig. 1 is a schematic flow chart of a parallel digital multi-beam synthesis method based on a GPU computing platform, as shown in fig. 1, the method includes the following steps:
and 100, transmitting a plurality of groups of beam coefficients and multiple paths of antenna data from a Central Processing Unit (CPU) to the GPU.
Step 101, in the GPU, digital multi-beam synthesis is performed based on the multiple sets of beam coefficients and the multiple paths of antenna data.
And 102, transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage.
In particular, the beam coefficients refer to parameters for digital beam synthesis, which may also be referred to as beam synthesis coefficients, weighting factors, weighting coefficients, etc.
The multi-antenna data refers to data (e.g., astronomical observation data) respectively received by the multi-antennas.
When digital multi-beam synthesis is performed, multiple groups of beam coefficients and multiple paths of antenna data in a CPU (memory) can be transmitted to a GPU (video memory), digital multi-beam synthesis is performed in the GPU, after a result of the digital multi-beam synthesis is obtained in the GPU, the result is transmitted to the CPU for storage, and therefore parallel computing capacity of the GPU can be fully utilized, and computing of the digital multi-beam synthesis can be accelerated.
Optionally, after the multiple sets of beam coefficients and the multiple paths of antenna data are transmitted to the GPU, the multiple sets of beam coefficients and the multiple paths of antenna data may be subjected to normalization processing (for example, format conversion, reordering, and the like, without limitation, in particular), and then the data after normalization processing may be used to perform calculation of beam synthesis.
Alternatively, the GPU may be one or more.
According to the parallel digital multi-beam synthesis method based on the GPU computing platform, data for digital multi-beam synthesis are transmitted to the GPU to be subjected to parallel acceleration processing, and the obtained digital multi-beam synthesis result is transmitted back to the CPU to be stored, so that the beam synthesis time can be remarkably reduced, the computing efficiency under large-scale data is improved, and the real-time performance of astronomical observation is ensured.
Optionally, based on the multiple sets of beam coefficients and the multiple antenna data, performing digital multi-beam synthesis includes:
for the synthesis of any target beam in any target integration time period, acquiring a set of beam coefficients for synthesizing the target beam and a set of multi-path antenna data in each time unit contained in the target integration time period based on a plurality of sets of beam coefficients and multi-path antenna data;
respectively carrying out weighted summation on a group of multipath antenna data in each time unit based on a group of beam coefficients;
obtaining a power spectrum of the target beam in a corresponding time unit based on a weighted summation result corresponding to each time unit and a multipath antenna data autocorrelation result summation item corresponding to each time unit;
and accumulating the power spectrums of the target beams in each time unit to obtain a synthesis result of the target beams.
Specifically, the target integration period may refer to any one integration period, and the target beam may refer to any one beam that needs to be synthesized.
An integration period may be divided into a plurality of time units, each of which may be a period of time. For example, one integration period may be a duration of 60ms, which may be divided into 12 time units, each time unit having a duration of 5ms.
Taking the synthesis of a certain beam in a certain integration time period as an example, the power spectrum of the beam in each time unit of the integration time period (that is, the same-frequency point power spectrum, and the beams of different frequency points are different beams) can be calculated first, and then the power spectrums of the beam in each time unit of the integration time period are accumulated to obtain the synthesis result of the beam.
When calculating the power spectrum of the beam in a certain time unit, a group of beam coefficients for beam synthesis and a group of multipath antenna data in the time unit can be determined first, the number of beam coefficients contained in the group of beam coefficients is equal to the number of antenna paths, each beam coefficient is accumulated (namely weighted summation) after complex multiplication calculation is carried out on each path of antenna data, simultaneously, autocorrelation calculation is carried out on each path of antenna data, and each path of autocorrelation result is added to obtain an association term (namely multipath antenna data autocorrelation result accumulation term), and the power spectrum of the beam in the time unit can be calculated based on the weighted summation result and multipath antenna data autocorrelation result accumulation term.
Optionally, based on the weighted summation result corresponding to each time unit and the multi-path antenna data autocorrelation result summation item corresponding to each time unit, obtaining a power spectrum of the target beam in the corresponding time unit includes:
determining the power spectrum of the target beam in any target time unit based on the following formula:
where a represents the power spectrum of the target beam in the target time unit,representing the weighted sum result corresponding to the target time unit, < >>And representing a multi-channel antenna data autocorrelation result accumulation item corresponding to the target time unit.
For the followingWherein M represents the number of antenna paths, ">C i Weight result representing one path of antenna data with sequence number i, < >>And I i Representing a corresponding one of the beam coefficients and one of the antenna data, respectively. />The meaning of the parameters in the method is the same as that described above, and the description is omitted.
Fig. 2 is an implementation schematic diagram of a parallel digital multi-beam synthesis method based on a GPU computing platform according to the present invention, and as shown in fig. 2, the main flow includes:
1. and acquiring a plurality of groups of beam coefficients, transmitting the beam coefficients to a GPU video memory, and regularizing the beam coefficients in the GPU.
2. And acquiring each path of antenna data, transmitting the original antenna data to a GPU video memory, and regularizing the original data in the GPU.
3. And performing kernel grouping according to the antenna data, and parallelly realizing beam synthesis of a plurality of observation integration time data.
4. In the GPU, using kernel, according to a beam synthesis formula, firstly, carrying out complex multiplication calculation on antenna data and beam coefficients, then accumulating, simultaneously carrying out autocorrelation calculation on each path of antenna data, adding each path of autocorrelation result to obtain an association term, and then carrying out association term elimination on each beam, wherein the formula is as follows:
wherein A represents the power spectrum of a beam in a time unit, M represents the number of antenna paths, C i The weighted result of one path of antenna data with the sequence number of i in the time unit is shown,and I i Representing a corresponding one of the beam coefficients and one of the antenna data, respectively.
5. The kernel is used in the GPU to accumulate the power spectrum of the same frequency point of a beam in an integration period, and the formula is as follows:
wherein B represents the result of power spectrum accumulation, i.e., the result of beam synthesis; t represents the number of time units within one integration period, which may also be referred to as the accumulated integration time unit (may be simply referred to as the accumulated integration time); a is that j Representing the power spectrum in one time unit with the sequence number j.
6. And transmitting the obtained digital multi-beam synthesis result from the GPU video memory to the CPU memory, and storing the result.
Optionally, the GPU includes at least one kernel (kernel), each kernel being configured to perform digital multi-beam synthesis during an integration period based on multiple antenna data and multiple sets of beam coefficients during the integration period.
Specifically, since the observation is continuous in time sequence, the data in each time unit is of a certain specification, the data in a plurality of time units which are continuously received can be grouped by utilizing the multi-stream (stream) characteristic of the GPU, so that the kernel of the plurality of streams processes the data in a plurality of time units in parallel, thereby improving the efficiency of beam synthesis and ensuring the observation instantaneity.
For example, fig. 3 is a schematic diagram of a design of parallel digital multi-beam synthesis of antenna data in Kernel according to the present invention, as shown in fig. 3, kernel 1 and Kernel2 of a GPU respectively process multiple antenna data of two different integration time periods (i.e. C in the figure) 1 C 2 C 3 C 4 C 5 C 6 C 7 ……C M )。
Optionally, each kernel contains a plurality of thread blocks (blocks), and each block contains a plurality of threads (threads); wherein each thread is used to calculate the power spectrum of a beam in a time unit.
For example, taking the kernel of FIG. 3 as an example, where a kernel contains multiple blocks, each block contains multiple reads, each read can calculate a beam in a unit of timePower spectrum, i.e. performing C as described above i And a calculation of a. Then, power spectrum accumulation in a plurality of time units is realized through block, so that a beam synthesis result is obtained.
And each thread is used for calculating the power spectrum of one wave beam in one time unit, so that partial wave beam synthesis results can be calculated in each thread of each kernel and stored into a global variable of a video memory, and after the execution of the kernel is finished, all threads obtain final results which are also ordered, thereby improving the wave beam synthesis efficiency.
Optionally, the block division manner in each kernel is determined based on the number of time units and the number of frequency points in one integration period, and the thread division manner in each block is determined based on the number of time units and the number of beam forming of one frequency point in one integration period.
For example, when digital multi-beam synthesis is performed, kernel grouping can be performed according to the size of the original data volume, parallel beam synthesis can be implemented by using multiple kernels, parallel design is performed in each kernel, and grid, block and thread division of the kernels can be performed by combining information such as the number b of beam synthesis, the number p of frequency points, accumulated integration time T and the like.
Taking the kernel shown in fig. 3 as an example, in each kernel, block division is first performed according to the accumulated integration time 12 and the number of frequency points p, so as to design blocks (2, p). And then designing a thread (6, b) in each block according to the accumulated integration time 12 and the beam synthesis quantity b. The total number of simultaneous threads running in each kernel is 12×p×b, and each thread can process each beam synthesis of M (e.g., 192) antenna data. And finally, 12 times of accumulation of accumulated integration time are realized by using a kernel design to obtain a final result.
The internal design of the kernel is carried out based on the information of the beam synthesis quantity b, the frequency point quantity p, the accumulated integration time T and the like, so that the adaptability, the flexibility and the expandability of the GPU for digital multi-beam synthesis are improved, astronomical observation can be effectively facilitated, the calculation efficiency is improved, and the observation instantaneity is ensured.
Optionally, the method further comprises:
the method comprises the steps of transmitting a plurality of groups of beam coefficients and multipath antenna data from a CPU to a GPU, performing digital multi-beam synthesis, and transmitting a digital multi-beam synthesis result from the GPU to the CPU for storage, wherein the steps are taken as three stages of a pipeline, and the digital multi-beam synthesis is realized in a pipeline parallel processing mode.
Specifically, fig. 4 is a schematic diagram of a data processing pipeline provided in the present invention, and as shown in fig. 4, the digital multi-beam synthesis for each group mainly includes three stages: the first stage is to transfer data from the memory to the video memory; the second stage is to start kernel function to perform multi-beam synthesis; and the third stage is to transmit the beam forming result from the video memory to the memory and store. The three stages can achieve pipeline parallel processing, and the CPU-GPU data transmission and GPU calculation are parallel in an asynchronous transmission mode, so that the overall running time is shortened, the CPU-GPU data transmission bottleneck frequently faced by GPU parallel acceleration is avoided, and the astronomical observation data processing efficiency can be maximized.
The parallel digital multi-beam synthesis device based on the GPU computing platform provided by the invention is described below, and the parallel digital multi-beam synthesis device based on the GPU computing platform described below and the parallel digital multi-beam synthesis method based on the GPU computing platform described above can be correspondingly referred to each other.
Fig. 5 is a schematic structural diagram of a parallel digital multi-beam synthesizing device based on a GPU computing platform according to the present invention, and as shown in fig. 5, the device includes:
the first transmission module 500 is configured to transmit the multiple sets of beam coefficients and multiple paths of antenna data from the central processing unit CPU to the GPU;
the beam synthesis module 510 is configured to perform digital multi-beam synthesis in the GPU based on multiple sets of beam coefficients and multiple antenna data;
and the second transmission module 520 is configured to transmit the result of the digital multi-beam synthesis from the GPU to the CPU for storage.
Optionally, based on the multiple sets of beam coefficients and the multiple antenna data, performing digital multi-beam synthesis includes:
for the synthesis of any target beam in any target integration time period, acquiring a set of beam coefficients for synthesizing the target beam and a set of multi-path antenna data in each time unit contained in the target integration time period based on a plurality of sets of beam coefficients and multi-path antenna data;
respectively carrying out weighted summation on a group of multipath antenna data in each time unit based on a group of beam coefficients;
obtaining a power spectrum of the target beam in a corresponding time unit based on a weighted summation result corresponding to each time unit and a multipath antenna data autocorrelation result summation item corresponding to each time unit;
and accumulating the power spectrums of the target beams in each time unit to obtain a synthesis result of the target beams.
Optionally, based on the weighted summation result corresponding to each time unit and the multi-path antenna data autocorrelation result summation item corresponding to each time unit, obtaining a power spectrum of the target beam in the corresponding time unit includes:
determining the power spectrum of the target beam in any target time unit based on the following formula:
where a represents the power spectrum of the target beam in the target time unit,representing the weighted sum result corresponding to the target time unit, < >>And representing a multi-channel antenna data autocorrelation result accumulation item corresponding to the target time unit.
Optionally, the GPU includes at least one kernel, each kernel being configured to perform digital multi-beam synthesis during an integration period based on multiple antenna data and multiple sets of beam coefficients during the integration period.
Optionally, each kernel contains a plurality of thread blocks, and each block contains a plurality of thread;
wherein each thread is used to calculate the power spectrum of a beam in a time unit.
Optionally, the block division manner in each kernel is determined based on the number of time units and the number of frequency points in one integration period, and the thread division manner in each block is determined based on the number of time units and the number of beam forming of one frequency point in one integration period.
Optionally, the apparatus further comprises:
the pipeline processing module is used for realizing digital multi-beam synthesis by adopting a pipeline parallel processing mode in three stages of a pipeline by taking the steps of transmitting a plurality of groups of beam coefficients and multi-path antenna data from the CPU to the GPU, carrying out digital multi-beam synthesis and transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage.
It should be noted that, the device provided by the present invention can implement all the method steps implemented by the method embodiment and achieve the same technical effects, and the parts and beneficial effects that are the same as those of the method embodiment in the present embodiment are not described in detail herein.
Fig. 6 is a schematic structural diagram of an electronic device according to the present invention, as shown in fig. 6, the electronic device may include: processor 610, communication interface (Communications Interface) 620, memory 630, and communication bus 640, wherein processor 610, communication interface 620, and memory 630 communicate with each other via communication bus 640. Processor 610 may invoke logic instructions in memory 630 to perform any of the parallel digital multi-beam synthesis methods based on GPU computing platforms provided in the embodiments described above.
Further, the logic instructions in the memory 630 may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
It should be noted that, the electronic device provided by the present invention can implement all the method steps implemented by the method embodiments and achieve the same technical effects, and the details and beneficial effects of the same parts and advantages as those of the method embodiments in the present embodiment are not described in detail.
In another aspect, the present invention further provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor, is implemented to perform any of the parallel digital multi-beam synthesis methods based on the GPU computing platform provided in the above embodiments.
It should be noted that, the non-transitory computer readable storage medium provided by the present invention can implement all the method steps implemented by the method embodiments and achieve the same technical effects, and detailed descriptions of the same parts and beneficial effects as those of the method embodiments in this embodiment are omitted.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. A parallel digital multi-beam synthesis method based on a graphics processor GPU computing platform, comprising:
transmitting a plurality of groups of beam coefficients and multi-path antenna data from a Central Processing Unit (CPU) to a Graphic Processing Unit (GPU);
in the GPU, digital multi-beam synthesis is performed based on the plurality of groups of beam coefficients and the multi-path antenna data;
transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage;
the digital multi-beam synthesis based on the plurality of sets of beam coefficients and the multi-path antenna data includes:
for synthesis of any target beam in any target integration time period, acquiring a set of beam coefficients for synthesizing the target beam and a set of multi-path antenna data in each time unit contained in the target integration time period based on the plurality of sets of beam coefficients and the multi-path antenna data;
respectively carrying out weighted summation on a group of multipath antenna data in each time unit based on the group of beam coefficients;
obtaining a power spectrum of the target beam in a corresponding time unit based on the weighted summation result respectively corresponding to each time unit and the multipath antenna data autocorrelation result summation item respectively corresponding to each time unit;
and accumulating the power spectrums of the target beam in each time unit to obtain a synthesis result of the target beam.
2. The GPU-based parallel digital multi-beam synthesis method according to claim 1, wherein the obtaining the power spectrum of the target beam in the corresponding time unit based on the weighted summation result corresponding to each time unit and the multi-path antenna data autocorrelation result summation term corresponding to each time unit comprises:
determining a power spectrum of the target beam within any target time unit based on the following formula:
where A represents the power spectrum of the target beam over the target time unit,representing the weighted sum result corresponding to said target time unit,/->Multiple paths of antenna data autocorrelation result accumulation items corresponding to the target time unitsThe method comprises the steps of carrying out a first treatment on the surface of the M represents the number of antenna paths, ">C i The weighted result of one path of antenna data with the sequence number of i is represented,and I i Representing a corresponding one of the beam coefficients and one of the antenna data, respectively.
3. The GPU-based parallel digital multi-beam synthesis method according to claim 1 or 2, wherein the GPU comprises at least one kernel, each kernel being configured to perform digital multi-beam synthesis during an integration period based on multiple antenna data and multiple sets of beam coefficients during the integration period.
4. The GPU-based parallel digital multi-beam synthesis method of claim 3, wherein each kernel comprises a plurality of thread blocks, each block comprising a plurality of thread;
wherein each said thread is used to calculate the power spectrum of a beam in a time unit.
5. The GPU-based parallel digital multi-beam synthesis method of claim 4, wherein the block partitioning in each kernel is determined based on the number of time units and the number of frequency points in an integration period, and the thread partitioning in each block is determined based on the number of time units and the number of beamforms of a frequency point in an integration period.
6. The GPU-based parallel digital multi-beam synthesis method according to any of claims 1-2, wherein the method further comprises:
and the step of transmitting the plurality of groups of beam coefficients and the plurality of paths of antenna data from the CPU to the GPU, the step of performing digital multi-beam synthesis and the step of transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage are taken as three stages of a pipeline, and the digital multi-beam synthesis is realized by adopting a pipeline parallel processing mode.
7. A parallel digital multi-beam synthesis apparatus based on a graphics processor GPU computing platform, comprising:
the first transmission module is used for transmitting a plurality of groups of beam coefficients and multiple paths of antenna data from a Central Processing Unit (CPU) to the GPU;
the beam synthesis module is used for carrying out digital multi-beam synthesis in the GPU based on the plurality of groups of beam coefficients and the multi-path antenna data;
the second transmission module is used for transmitting the digital multi-beam synthesis result from the GPU to the CPU for storage;
the digital multi-beam synthesis based on the plurality of sets of beam coefficients and the multi-path antenna data includes:
for synthesis of any target beam in any target integration time period, acquiring a set of beam coefficients for synthesizing the target beam and a set of multi-path antenna data in each time unit contained in the target integration time period based on the plurality of sets of beam coefficients and the multi-path antenna data;
respectively carrying out weighted summation on a group of multipath antenna data in each time unit based on the group of beam coefficients;
obtaining a power spectrum of the target beam in a corresponding time unit based on the weighted summation result respectively corresponding to each time unit and the multipath antenna data autocorrelation result summation item respectively corresponding to each time unit;
and accumulating the power spectrums of the target beam in each time unit to obtain a synthesis result of the target beam.
8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the parallel digital multi-beam synthesis method based on a GPU computing platform as claimed in any of claims 1 to 6 when the program is executed by the processor.
9. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the parallel digital multi-beam synthesis method based on a GPU computing platform as claimed in any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310226191.2A CN116418381B (en) | 2023-03-02 | 2023-03-02 | Parallel digital multi-beam synthesis method and device based on GPU computing platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310226191.2A CN116418381B (en) | 2023-03-02 | 2023-03-02 | Parallel digital multi-beam synthesis method and device based on GPU computing platform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116418381A CN116418381A (en) | 2023-07-11 |
| CN116418381B true CN116418381B (en) | 2024-02-13 |
Family
ID=87055670
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310226191.2A Active CN116418381B (en) | 2023-03-02 | 2023-03-02 | Parallel digital multi-beam synthesis method and device based on GPU computing platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116418381B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106299702A (en) * | 2016-09-14 | 2017-01-04 | 中国科学院上海天文台 | A kind of low frequency radio array digital bea mforming system and method |
| CN107017933A (en) * | 2017-05-09 | 2017-08-04 | 电子科技大学 | A kind of MIMO data transmission methods and device for merging smart antenna |
| US9791552B1 (en) * | 2014-11-19 | 2017-10-17 | Src, Inc. | On-site calibration of array antenna systems |
| CN111426889A (en) * | 2020-04-14 | 2020-07-17 | 中国科学院国家天文台 | Broadband dual-mode digital receiver and signal processing method thereof |
| CN113162670A (en) * | 2021-03-30 | 2021-07-23 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Digital multi-beam correction and synthesis method |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1118201C (en) * | 1999-08-11 | 2003-08-13 | 信息产业部电信科学技术研究院 | Interference counteracting method based on intelligent antenna |
| KR100383669B1 (en) * | 2000-10-12 | 2003-05-14 | 한국전자통신연구원 | A Method and Apparatus for Code Acquisition in a CDMA Basestation System Using an Array Antenna |
| US10863582B2 (en) * | 2018-02-16 | 2020-12-08 | Apple Inc. | Methods to signal antenna panel capability of user equipment (UE) for carrier aggregation (CA) in millimeter-wave (MMWAVE) frequency bands |
-
2023
- 2023-03-02 CN CN202310226191.2A patent/CN116418381B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9791552B1 (en) * | 2014-11-19 | 2017-10-17 | Src, Inc. | On-site calibration of array antenna systems |
| CN106299702A (en) * | 2016-09-14 | 2017-01-04 | 中国科学院上海天文台 | A kind of low frequency radio array digital bea mforming system and method |
| CN107017933A (en) * | 2017-05-09 | 2017-08-04 | 电子科技大学 | A kind of MIMO data transmission methods and device for merging smart antenna |
| CN111426889A (en) * | 2020-04-14 | 2020-07-17 | 中国科学院国家天文台 | Broadband dual-mode digital receiver and signal processing method thereof |
| CN113162670A (en) * | 2021-03-30 | 2021-07-23 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Digital multi-beam correction and synthesis method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116418381A (en) | 2023-07-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108462521B (en) | Anti-interference realization method of self-adaptive array antenna | |
| CA2206194C (en) | Method and system for digital beam forming | |
| CN104991236B (en) | A kind of single base MIMO radar not rounded signal coherence source Wave arrival direction estimating method | |
| CN113222822B (en) | Hyperspectral image super-resolution reconstruction method based on multi-scale transformation | |
| CN111461311B (en) | Convolutional neural network operation acceleration method and device based on many-core processor | |
| CN106680815B (en) | MIMO radar imaging method based on tensor sparse representation | |
| CN102680945B (en) | Doppler modulation frequency estimation method based on field programmable gate array (FPGA) | |
| CN113156383B (en) | FPGA-based radar signal anti-interference method, system and data processing equipment | |
| CN104656073B (en) | Three-dimensional imaging sonar wave beam forming method and implementation method on multi-core processor | |
| CN108802726B (en) | Synthetic aperture radar imaging method based on Graphic Processing Unit (GPU) | |
| CN106526565A (en) | Single-bit spatial spectrum estimation method based on support vector machine | |
| CN109946663B (en) | Linear complexity Massive MIMO target space orientation estimation method and device | |
| CN116418381B (en) | Parallel digital multi-beam synthesis method and device based on GPU computing platform | |
| CN109188346B (en) | Single snapshot DOA estimation method for large-scale uniform cylindrical array | |
| CN111552559B (en) | Broadband signal DDC system design method based on GPU | |
| CN106019334B (en) | A kind of correlator implementation method of the navigation software receiver based on GPU | |
| Barhen et al. | High performance FFT on multicore processors | |
| CN108680892B (en) | Estimation method of direction of arrival of co-prime array based on angle-space frequency domain fast Fourier transform | |
| CN111865385B (en) | Two-dimensional planar array digital beam forming method based on FPGA | |
| CN110471041B (en) | Vivado HLS-based two-dimensional DOA estimation method | |
| CN107135026A (en) | Robust ada- ptive beamformer method based on matrix reconstruction in the presence of unknown mutual coupling | |
| CN114397619A (en) | Two-dimensional positioning algorithm based on non-uniform sparse array | |
| CN114244460B (en) | Heterogeneous accelerated multi-path channel signal real-time generation method | |
| CN113671477B (en) | Radar target distance estimation method based on graph signal processing | |
| JP7058810B2 (en) | Signal processing system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |