Chirp信号参数估计算法性能比较

通过数值仿真定量地比较了三种Chirp信号参数估计算法——解线调法、迭代估计法和局部搜索最大似然法的性能,并定性地比较了算法的运算量。仿真结果表明,在三种算法中,局部搜索最大似然法的估计性能最好,而运算量居中;解线调法运算量最大,但估计性能居中;迭代估计法的估计性能最差,但运算量最小。对于实际系统,应根据不同的估计精度和运算量要求,灵活选择不同的算法。综合考虑估计性能和运算量之间的折衷可以得到结论,在三种算法中局部搜索最大似然法是一种相对较好的选择。

Performance comparison of parameters estimation algorithms of chirp signals

In this paper,a quantitive performance comparison is made by numerical experiments among three different parameters estimation algorithms for chirp signals: dechirp method,iterative approach and local-searching maximum-likelihood(LSML) method.Also their computation loads are compared qualitatively.The results of simulation show that the estimation performance of LSML is the best and its computation loads are medial.The computation loads of dechirp method is the highest but its estimation performance is medial.The estimation performance of iterative approach is the worst;however,its computation loads is the lowest.For practical systems,we should make a flexible choice among different algorithms according to different requirements for estimation accuracy and computation loads.Considering both the estimation performance and the computation loads,we conclude that LSML method is a relatively good choice in the three algorithms.