CN110320501A - A kind of radar signal impulse compression method based on GPU - Google Patents
A kind of radar signal impulse compression method based on GPU Download PDFInfo
- Publication number
- CN110320501A CN110320501A CN201910570816.0A CN201910570816A CN110320501A CN 110320501 A CN110320501 A CN 110320501A CN 201910570816 A CN201910570816 A CN 201910570816A CN 110320501 A CN110320501 A CN 110320501A
- Authority
- CN
- China
- Prior art keywords
- gpu
- data
- signal
- pulse
- fft
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/28—Details of pulse systems
- G01S7/285—Receivers
- G01S7/295—Means for transforming co-ordinates or for evaluating data, e.g. using computers
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention belongs to Radar Signal Processing Technology fields, are related to a kind of radar signal impulse compression method based on GPU.The present invention carry out wait pulse compression signal FFT and pulse pressure after signal IFFT when, can be with parallel processing mass data using the FFT interface in GPU；When the frequency-region signal of pulse pressure and matched filter being waited to carry out dot product, data vector is carried out by segment processing using the parallel optimization of GPU.Calculative data are divided into multistage, the multiplication of every segment data gives different GPU core processing, and the data being segmented realize parallel processing；In the case where GPU video memory is sufficient, the compression of signal pulse of multiple-pulse multichannel can be supported to handle, and kept for operation time meet requirement of real-time.
Description
Technical field
The invention belongs to Radar Signal Processing Technology fields, are related to a kind of radar signal impulse compression method based on GPU.
Background technique
Conventional radar systems are substantially exploitation mould by " centered on hardware design " and " towards special function " two aspects
Formula needs to match different hardware and softwares for the radar of different model, and hardware and software is upgraded very big limit again
System, the reconstruct of function and the update of technology are difficult to realize so as to cause conventional radar.Software implementation radar uses level decoupling
Design philosophy, by formulate series of standardsization and standardization component, data standard, by radar system be divided into several have compared with
The level of strong independence, developer can be to the function parallelization exploitations of different levels in system, to quickly push and respond system
The technology innovation of system, the speed for improving expanding of system function and performance boost.Software implementation radar has standardization, digitlization and mould
The big feature of blockization three, the development mode for being " using software technology as core, application-oriented demand " and be based on opening system frame
The radar system of structure.
GPU full name is graphics processor (Graphics Processor Unit), it is the core cell of video card.From GPU
The rendering task to figure is served as since generation always, as NVIDIA company is proposed unified computing architecture
(Computer Unified Device Architecture, CUDA), uses C language as development language, the programming of GPU
To become convenient, simple, GPU is also promoted to yield unusually brilliant results in the data calculation processing field other than graphics calculations, such as
General high-performance computing sector.In structure, it is substantially controller and cache register on CPU, and possesses on GPU architecture big
The logical unit of amount, it is more suitable that this allows for the GPU in parallel processing mass data.And the exploitation high-performance meter of GPU
Calculation ability is considerably beyond CPU, and NVIDAGP company's T esla K80 single precision computing capability has reached 5.6TFLOPS (often
Second flops, FLOPS).
Radar when carrying out ranging processing, is realized using the delay time of target echo.Therefore, radar
Distance resolution is related to pulsewidth, i.e., pulsewidth is shorter, and the resolution ratio of radar is stronger.But if shortening the pulse width of radar,
The transmitting mean power of radar will be reduced, is shortened so as to cause the operating distance of radar, pulse compression technique is exactly to pass through thunder
Up to time-bandwidth product big signal when transmitting, guarantee the distance resolution that radar will not be reduced when improving radar data reduction.
Summary of the invention
In order to meet the requirement of real-time of Radar Signal Processing, pulse compression system must have very high time performance,
Present invention utilizes the high-speed computational capability of GPU, devises a kind of pulse accelerated based on GPU and compress implementation method, it can be with
Faster speed completes the pulse compression of signal, meets the requirement of system signal processing real-time.
The technical scheme is that a kind of Radar Signal In-Pulse Characteristics punching press contracting implementation method accelerated based on GPU, feature are existed
In, comprising the following steps:
S1, data are respectively put into GPU video memory, GPU thread structure is three dimensional network structure, enables each x of three-dimensional grid
Dimension all thread blocks in direction complete the calculating in a corresponding channel, the direction grid y dimension is port number, grid z-dimension is arteries and veins
Rush number；
S2, treat Signal for Pulse carry out Fast Fourier Transform (FFT), using GPU on highly-parallel optimization FFT interface,
Multiple computing units, a part of Fast Fourier Transform (FFT) of parallel computation on each computing unit, finally by institute are utilized in GPU
The FFT for having computing unit to complete carries out concatenation, obtains the frequency-region signal to Signal for Pulse；
S3, the dot product for obtaining Signal for Pulse and radar reference signal: GPU is realized by the way that data vector is carried out segmentation
Parallel optimization, calculative data are divided into multistage first, the multiplication calculating of every segment data gives different GPU core to count
It calculates, realizes the parallel processing between data.In GPU, the multiplication that different threads carries out different data section fastly is calculated, per thread
Different threads between fast are completed different and data point multiplication and are calculated.
S4, Inverse Fast Fourier Transforms are carried out to vector dot result, same use highly-parallel on GPU optimizes IFFT
Interface.
Beneficial effects of the present invention are to carry out parallel optimization to algorithm, and the high-speed parallel for enabling to adapt to GPU calculates energy
Power；The pulse compression to radar signal is realized, guarantees to possess preferable Range resolution while improving the signal-to-noise ratio of signal
Rate；Using GPU parallel processing, meet requirement of real-time；In the case where GPU video memory is sufficient, multiple-pulse multichannel can be supported
Compression of signal pulse processing, and kept for operation time meet requirement of real-time.
Detailed description of the invention
Fig. 1 is that frequency-domain impulse compression algorithm realizes schematic diagram；
Fig. 2 is dot product optimization schematic diagram；
Fig. 3 is flow diagram of the invention；
Fig. 4 is that the FFT performance of GPU and CPU compares；
Fig. 5 is target echo result schematic diagram (a) CPU echo result；(b) GPU echo result；
Fig. 6 is GPU and target cpu echo resultant error schematic diagram.
Specific embodiment
Further technical solution of the present invention is described with reference to the accompanying drawing.
As shown in Figure 1, technical solution of the present invention includes the following steps:
Implementation method is compressed based on the GPU pulse accelerated, comprising the following steps:
Step 1: placing data into GPU video memory, GPU thread structure is three-dimensional grid.
The each X of grid ties up the calculating that all threads complete a corresponding channel fastly；
Grid y direction dimension is port number；
Grid z direction dimension is umber of pulse, to reach while carry out the pulse compression of multiple-pulse multi channel signals；
Step 2: the FFT interface optimized using highly-parallel on GPU carries out fast Fourier fortune to waiting Signal for Pulse
It calculates；
Step 3: complete wait Signal for Pulse and radar reference signal vector dot: on GPU parallel optimization by pair
Data vector is segmented to realize.Calculative data are divided into multistage, the multiplication calculating of every segment data is given different
GPU core calculates, and realizes the parallel processing being segmented between data.Meanwhile the calculating of the different data point in the same data segment
It also can be realized parallel in GPU, in cuda model, the threads in each block completes the multiplication of different data point；
Become step 4: carrying out inverse Fourier to vector dot result using the IFFT interface that highly-parallel optimizes on GPU
It changes, final result can be obtained.
Fig. 4 is that the FFT performance of GPU and CPU compares, it can be seen that when FFT points are very big, GPU completes FFT and executes
Speed is substantially better than CPU.
Fig. 5 is pulse compression technology result verification of the invention, can be obtained by comparison, and method of the invention is available correctly
As a result.
Fig. 6 is GPU and target cpu echo resultant error schematic diagram, can obtain GPU echo calculated result and CPU simulation result
Error very little, so explanation realizes that pulse pressure is verified on GPU.
Pulse is compressed on CPU and GPU time-consuming comparison as shown in table 1 below,
Table 1:CPU and GPU complete the comparison of pulse compression time
, it is apparent that realizing that the pulse compression time-consuming of echo-signal is significantly less than using CPU using GPU, performance is excellent
It is good.
Claims (1)
1. a kind of radar signal impulse compression method based on GPU characterized by comprising
S1, data are respectively put into GPU video memory, GPU thread structure is three dimensional network structure, enables each direction x of three-dimensional grid
All thread blocks of dimension complete the calculating in a corresponding channel, the direction grid y dimension is port number, grid z-dimension is umber of pulse；
S2, treat pulse compression signal carry out Fast Fourier Transform (FFT), using the FFT interface of GPU, in GPU utilize multiple meters
Unit is calculated, a part of Fast Fourier Transform (FFT) of parallel computation on each computing unit finally completes all computing units
FFT carry out concatenation, obtain the frequency-region signal to pulse compression signal；
S3, the dot product for obtaining pulse compression signal and radar reference signal: GPU is realized by the way that data vector is carried out segmentation
Parallel optimization, calculative data are divided into multistage first, the multiplication calculating of every segment data gives different GPU core to count
It calculates, realizes the parallel processing between data；In GPU, the multiplication that different threads carries out different data section fastly is calculated, per thread
Different threads between fast are completed different and data point multiplication and are calculated；
S4, Inverse Fast Fourier Transforms, the same IFFT interface using in GPU are carried out to vector dot result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910570816.0A CN110320501A (en) | 2019-06-27 | 2019-06-27 | A kind of radar signal impulse compression method based on GPU |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910570816.0A CN110320501A (en) | 2019-06-27 | 2019-06-27 | A kind of radar signal impulse compression method based on GPU |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110320501A true CN110320501A (en) | 2019-10-11 |
Family
ID=68121389
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910570816.0A Pending CN110320501A (en) | 2019-06-27 | 2019-06-27 | A kind of radar signal impulse compression method based on GPU |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110320501A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114690127A (en) * | 2022-03-24 | 2022-07-01 | 西安电子科技大学 | FT-M6678-based real-time pulse pressure method for large-time wide-bandwidth signals |
| CN117331512A (en) * | 2023-12-01 | 2024-01-02 | 芯动微电子科技(武汉)有限公司 | Data compression and processing method for executing write operation on GPU (graphics processing unit) nuclear memory |
| CN118432583A (en) * | 2024-07-04 | 2024-08-02 | 成都蓝色起源科技有限公司 | Linear filtering method and device based on overlap-add method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104181510A (en) * | 2014-09-05 | 2014-12-03 | 电子科技大学 | Parallel radar pulse compression method |
| CN105137402A (en) * | 2015-08-19 | 2015-12-09 | 电子科技大学 | Airborne foresight scanning radar parallel processing method based on GPU |
| CN107153547A (en) * | 2017-05-15 | 2017-09-12 | 电子科技大学 | A kind of software implementation radar storehouse mode signal handles middleware |
| CN107315168A (en) * | 2017-07-11 | 2017-11-03 | 电子科技大学 | A kind of software implementation radar signal data handling system and method |
| CN107728114A (en) * | 2017-11-14 | 2018-02-23 | 四川泰立科技股份有限公司 | Slower-velocity target surveillance radar signal processing method based on GPU |
-
2019
- 2019-06-27 CN CN201910570816.0A patent/CN110320501A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104181510A (en) * | 2014-09-05 | 2014-12-03 | 电子科技大学 | Parallel radar pulse compression method |
| CN105137402A (en) * | 2015-08-19 | 2015-12-09 | 电子科技大学 | Airborne foresight scanning radar parallel processing method based on GPU |
| CN107153547A (en) * | 2017-05-15 | 2017-09-12 | 电子科技大学 | A kind of software implementation radar storehouse mode signal handles middleware |
| CN107315168A (en) * | 2017-07-11 | 2017-11-03 | 电子科技大学 | A kind of software implementation radar signal data handling system and method |
| CN107728114A (en) * | 2017-11-14 | 2018-02-23 | 四川泰立科技股份有限公司 | Slower-velocity target surveillance radar signal processing method based on GPU |
Non-Patent Citations (2)
| Title |
|---|
| 周晓露: "基于GPU的阵列雷达信号处理及实现", 《中国优秀硕士学位论文全文数据库》 * |
| 彭培: "基于CPU/GPU处理器的雷达脉冲压缩算法并行机制研究", 《舰船电子工程》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114690127A (en) * | 2022-03-24 | 2022-07-01 | 西安电子科技大学 | FT-M6678-based real-time pulse pressure method for large-time wide-bandwidth signals |
| CN114690127B (en) * | 2022-03-24 | 2024-07-16 | 西安电子科技大学 | FT-M6678-based real-time pulse pressure method for large-time wide bandwidth signals |
| CN117331512A (en) * | 2023-12-01 | 2024-01-02 | 芯动微电子科技(武汉)有限公司 | Data compression and processing method for executing write operation on GPU (graphics processing unit) nuclear memory |
| CN117331512B (en) * | 2023-12-01 | 2024-04-12 | 芯动微电子科技(武汉)有限公司 | Data compression and processing method for executing write operation on GPU (graphics processing unit) nuclear memory |
| CN118432583A (en) * | 2024-07-04 | 2024-08-02 | 成都蓝色起源科技有限公司 | Linear filtering method and device based on overlap-add method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110320501A (en) | A kind of radar signal impulse compression method based on GPU | |
| CN103197317B (en) | Synthetic aperture radar (SAR) imaging method based on field programmable gate array (FPGA) | |
| CN102680945B (en) | Doppler modulation frequency estimation method based on field programmable gate array (FPGA) | |
| CN102520390A (en) | Two-dimensional MUSIC (multiple signal classification) direction-detecting device for uniform circular array | |
| CN110361691A (en) | Coherent DOA based on nonuniform noise estimates FPGA implementation method | |
| CN103901405B (en) | Block floating point frequency domain four road pulse shortener and impulse compression methods thereof in real time | |
| CN106291473A (en) | Nested type aerial array method to set up | |
| CN102567282B (en) | In general dsp processor, FFT calculates implement device and method | |
| CN105182292B (en) | A kind of more waveform phase coding methods based on pattern search algorithm | |
| CN108614234B (en) | Direction-of-arrival estimation method based on multi-sampling snapshot co-prime array received signal fast Fourier inverse transformation | |
| CN108508426B (en) | SAR echo signal generation method based on multi-core DSP and echo simulator | |
| CN103776907A (en) | Ultrasonic phased array received signal fine delaying method based on sinc interpolation | |
| CN103728616A (en) | Field programmable gate array (FPGA) based inverse synthetic aperture radar (ISAP) imaging parallel envelope alignment method | |
| CN104731762A (en) | Cubic phase signal parameter estimation method based on cyclic shift | |
| Wang et al. | Energy and performance characterization of mobile heterogeneous computing | |
| CN111337896A (en) | Method for realizing moving target detection acceleration | |
| CN116561541A (en) | Random dynamic load identification model, training method of model and identification method | |
| CN109782245A (en) | Wave arrival direction estimating method and device, radar, readable storage medium storing program for executing | |
| CN108415005A (en) | A kind of passive location delay time estimation method and device | |
| CN108680892A (en) | Relatively prime array Wave arrival direction estimating method based on angle-spatial frequency domain Fast Fourier Transform (FFT) | |
| CN110736970B (en) | Radar target rapid identification method based on ASIC machine learning processor | |
| CN109921800B (en) | Frequency domain pulse compression method based on self-customized floating point format | |
| CN103926567B (en) | high-speed real-time pulse compression algorithm | |
| CN105785348A (en) | Sonar signal processing method based on ZYNQ-7000 platform | |
| CN112597432A (en) | Method and system for realizing acceleration of complex sequence cross-correlation on FPGA (field programmable Gate array) based on FFT (fast Fourier transform) algorithm |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191011 |
|
| WD01 | Invention patent application deemed withdrawn after publication |