EM-ACO algorithm and its application to parameters estimate of multiple ultrasonic echoes
ZHOU Fang1,2， ZHANG Xiao-feng2*
(1 Institute of Applied Acoustics, Shaanxi Normal University, Xi′an 710062, Shaanxi，China;2 College of Information Science and Technology, Shihezi University, Shihezi 832003, Xinjiang, China)
Abstract:
Aiming at the defects that the convergence speed is so slow and the iterative results depend on the initial values seriously in the application of EM algorithm estimated the parameters of multiple ultrasonic echoes, a new method for parameters estimation of multiple ultrasonic echoes: EM-ACO algorithm is proposed, which combines the advantages of ant colony algorithm and EM algorithm. The new algorithm can not only obtain the good results at different initial guesses and improves the convergence speed of EM algorithm significantly, but also achieve a higher precision. According to Gaussian Echoes model, this new algorithm is applied to the parameters estimation of multiple ultrasonic echoes for different signal to noise ratio(SNRs). The simulation results show that EM-ACO algorithm can successfully estimate the parameters of multiple ultrasonic echoes with fewer iterations and has a higher precision in conditions of all sorts of different initial values.
KeyWords:
EM algorithm; ant colony algorithm; parameters estimation; gaussian echo model; multiple ultrasonic echoes
