Font Size: a A A

Formation Processor GPU-based Adaptive Beam

Posted on:2015-03-21Degree:MasterType:Thesis
Country:ChinaCandidate:X YangFull Text:PDF
GTID:2268330425988069Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
For adaptive beamforming processor in radar system, how to implement high-speed parallel adaptive calculation is an important problem.In order to solve this problem,we explore how to use the parallel processing architecture of graphics processor (GPU), and retrofit existing adaptive beamforming algorithm to improve adaptive beamforming processor real-time processing capabilities.The main work of this paper include:(1) Analyzed the hardware architecture of graphics processor (GPU) and the hardware, software framework for parallel processing.Introduced the characteristics of general purpose parallel computing architecture:CUD A (Compute Unified Device Architecture). Describes the software development process and tried the parallel processing of algorithm in the hardware and software environment of CUD A.(2) According to the features of GPU and CUDA,we transformed the adaptive beam forming algorithm based on orthogonal projection process, to enable them to parallel computing in the hardware platform, completed optimized programming algorithms, and completed the evaluation of the performance of the algorithm and tested on the hardware platform. Compared with the DSP,experimental results show that for the same computational task, using the improved orthogonal projection algorithm based on GPU, the calculation speed more than10times.(3) Transformed the Fast Reduced Rank Minimum Variance Beamfoemer,to enable them to parallel computing in the hardware platforms of GPU and CUDA,and completed the evaluation of the performance of the algorithm and tested on the hardware platform.Compared with the DSP.experimental results show that for the same computational task, using the improved FRRMVB algorithm based on GPU, the calculation speed more than1times.(4) Proposed a scheme of implementing the adaptive beamforming processer based on multi GPU+FPGA, provided an integrated frame diagram, a detailed introduction of the components of it. main feature and interfaces. In the end. provided a detailed analyze of data flow in the adaptive beamforming processer.
Keywords/Search Tags:GPU, Parallel Computing, Adaptive Beamforming Processor
PDF Full Text Request
Related items