Dynamic time programming based on ant colony algorithm

The random time changing behavior is very popular in speech signal. In order to correct it the warping method is often used in speech signal processing which based on template matching. Ant colony algorithm is a novel random optimization algorithm. It had shown many promising properties in solving complicated optimization problems. Applying the thought of ant colony algorithm to speech signal processing this paper presents a new dynamic time programming based on ant colony algorithm - ADTP. It uses both the global and the local characters of speech signal. The theoretic analysis and simulation experiments all certify the new algorithm feasibility. The matching results of the new method can show more accurate similarity between speech signals than the DTW method.