Finite impulse response(FIR)filters with the advantages of linear phase and stability are widely used in radar signal processing,image processing and electronic countermeasures.However,high performance FIR filters usually require higher order,which leads to high hardware implementation cost and high system power consumption.To solve this problem,an improved iter-ative reweighted orthogonal matching tracking algorithm is proposed to design sparse linear phase FIR filter.The scheme employs the Frobenius norm to simplify searching for the desired sparse FIR filter at each iteration,and a log function is introduced to update the weights iteratively.The simulation results show that the filters designed by the proposed scheme has shorter effective fil-ter length and fewer non-zero filter taps than the existing sparse filter design methods in some cases,which avoids the heavy use of multipliers and adders in hardware implementation,thus ef-fectively reducing the implementation cost and system power consumption.
关键词
稀疏FIR滤波器/线性相位/正交匹配追踪/Frobenius范数
Key words
sparse FIR filter/linear phase/orthogonal matching pursuit/Frobenius norm
维普
万方数据