Unlock instant, AI-driven research and patent intelligence for your innovation.

Airborne Radar Motion Velocity Estimation Method Based on Accumulation and Parallelization

An airborne radar and motion speed technology, applied in the direction of calculation, program control design, multi-program device, etc., can solve the problems of insufficient use of hardware computing units and slow computing speed, so as to improve the utilization rate of hardware resources, improve the cumulative and The effect of reducing operation speed and algorithm complexity

Active Publication Date: 2021-07-27
XIDIAN UNIV
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem that the accumulative sum used in the airborne synthetic aperture radar in the prior art is not fast enough to calculate the motion velocity vector and the use of hardware computing units is insufficient, the present invention proposes an accumulative sum parallel algorithm with a significant acceleration effect and high resource utilization. Optimized Airborne Radar Motion Velocity Estimation Method

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Airborne Radar Motion Velocity Estimation Method Based on Accumulation and Parallelization
  • Airborne Radar Motion Velocity Estimation Method Based on Accumulation and Parallelization
  • Airborne Radar Motion Velocity Estimation Method Based on Accumulation and Parallelization

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0022] During the imaging process of airborne synthetic aperture radar, due to the influence of airflow, the motion instability is very large. If motion compensation is not adopted, the recorded data will be greatly distorted due to the influence of unstable factors, which will reduce the imaging quality. Can't even image. Before motion compensation, the estimation of motion parameters is essential, and if the acceleration vector estimated in the previous stage is to be time-integrated to obtain the velocity vector for real-time imaging on the machine, fast accumulation and solution are required. However, the current accumulation and serial Row-solve methods fall short of this speed requirement. For this reason, the present invention is through innovation, proposes the airborne radar motion speed estimation method based on accumulation and parallelization, see figure 1 , including the following steps:

[0023] Step 1 Generate the motion acceleration vector of the airborne ra...

Embodiment 2

[0030] The airborne radar speed estimation method based on accumulation and parallelization is the same as embodiment 1, see figure 2 , figure 2 For the thread configuration schematic diagram of the GPU kernel function of the present invention, the process of the thread configuration of the design GPU kernel function described in step 2 includes:

[0031] 2a) Analysis of hardware resource capacity: limited by hardware resources, there is an upper limit on the number of thread blocks that can be opened up by the GPU and the size of each thread block. The maximum number of thread blocks that can be opened up by the GPU grid is GridMax. The maximum number of threads is BlockMax. In this example, GridMax=2147483647*65535*65535, BlockMax=1024.

[0032] 2b) Analyze the vector to be calculated, and determine the calculation principle: the length of the airborne radar motion acceleration vector A is l, l=2 n , n is a positive integer, calculation principle: in the calculation, ea...

Embodiment 3

[0039] The airborne radar speed estimation method based on accumulation and parallelization is the same as that in Embodiment 1-2, and the process of generating the airborne radar motion speed variation vector described in step 3 includes:

[0040] 3a) Allocate GPU shared memory: according to the thread configuration of the GPU kernel function, allocate GPU shared memory for each thread block, the size of the shared memory in each thread block is 2*BlockNum*sizeof(type), a total of GridNum thread blocks, type It is the data type of airborne radar motion parameters, and sizeof(type) indicates the byte length occupied by a variable of this data type. In this example, GridNum=16, BlockNum=1024, the data type of the radar motion parameter is single-precision floating point (float), and sizeof(type)=4.

[0041] 3b) Generate airborne radar motion velocity variation vector: Under the thread configuration of the GPU kernel function, according to the calculation principle in step 2, en...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an airborne radar motion speed estimation method based on accumulation and parallelization, which solves the problems of low calculation efficiency and insufficient utilization of hardware resources in the airborne radar motion speed estimation. The implementation steps include: generating the motion acceleration vector of the airborne radar in the GPU; designing the thread configuration of the GPU kernel function; generating the motion velocity variation vector of the airborne radar; parallel accumulation and calculation in the thread block to obtain the airborne radar in each thread block of the GPU Temporary motion velocity vector; parallel accumulation and calculation between GPU blocks to obtain the airborne radar motion velocity vector. According to the hardware resource capacity combined with the initial data volume, the invention sets reasonable thread configuration; builds an accumulation and solution model suitable for a parallel computing platform; and performs parallel computing on data in thread blocks and between thread blocks. The complexity of the algorithm is reduced, the operating efficiency is improved, the utilization of hardware resources is more reasonable and effective, and the overall data processing speed and utilization of hardware resources are improved. Airborne radar fast-look imaging for SAR imaging.

Description

technical field [0001] The invention belongs to the technical field of radar data processing, and relates to GPU parallel acceleration technology, in particular to an airborne radar motion speed estimation method based on accumulation and parallelization, which is used for airborne radar motion speed estimation in SAR imaging. Background technique [0002] In airborne synthetic aperture radar speed estimation, the radar motion velocity vector is obtained by integrating the radar motion acceleration vector mainly by the cumulative sum method. [0003] The traditional serial method based on the CPU platform is to superimpose the elements in the vector one by one in order. Although this method is simple and easy to implement, because the number of operations of this method is proportional to the length of the vector, the algorithm complexity is high, so the performance is relatively common. It will take a long time to execute this method on the CPU, which affects the real-time ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F9/50G06F9/54
CPCG06F9/5027G06F9/544G06F2209/5018
Inventor 李良超薛心竹郭亮杨军
Owner XIDIAN UNIV