Implementation Of Adaptive Filter Based On Dynamic Searching Particle Swarm Optimization

Adaptive filter,which is an important part of modern digital signal processing,is able to adjust its weight coefficients automatically in the filtering process according to the characteristics of the received input signals,so as to filter,smooth and predict the received signals effectively.As we all known,the classical adaptive filtering algorithms are slow to converge and incapable of meeting the requirements of high-performance realtime signal processing.In this paper,we summarize the problem of optimal value of adaptive filter as the multi-objective optimization problem,which could be solved by a powerful tool,Particle Swarm Optimization algorithm with the great advantages of fast convergence speed,strong global search ability.Therefore,this paper proposed a novel adaptive filtering algorithm based on Particle Swarm Optimization.After summarizing the previous related works,we theoretically analyze that the premise to search for the optimal solution by Particle Swarm Optimization is the search space is static.Therefore,in a wide-sense stationary environment,this paper proposed an adaptive filter design method based on standard Particle Swarm Optimization.Aiming to overcome the shortcoming of adaptive filtering algorithm based on standard Particle Swarm Optimization mentioned above,we proposed another novel adaptive filtering algorithm based on the Dynamic Searching Particle Swarm Optimization,which can work in a nonstationary environment.Some convergence analysis work of the proposed algorithm has also done.The simulation results show that the adaptive filtering algorithm based on the Dynamic Search Particle Swarm Optimization can not only converge quickly,but also has strong stability.Finally,a user-defined floating-point adaptive spatial filter for interfere nce suppression,using Dynamic Search Particle Swarm Optimization based on FPGA's pipelined polyphase architecture on FPGA,has been implemented.The pipelined polyphase signal processing technique enables large-scale Particle Swarm Optimiza tio n without reducing the performance of the algorithm,while greatly reducing hardware resources consumption.The proposed architectures have been developed in hardware description language using Verilog HDL and VHDL(Very High Speed Integrated Circuits Hardware Description Language).For convenience and simplification,a cosimulation technique by Modelsim and Matlab/Simulink with HDL Verifier is applied to verify the simulated results and the proposed architectures are synthesizable in the XILINX Vivado tool.