Mobile sound source time delay of arrival (TDOA) estimation method based on dynamic programming algorithm

Abstract

The invention discloses a mobile sound source time delay of arrival (TDOA) estimation method based on a dynamic programming algorithm. The method takes advantage of the coherence of mobile sound source time delay estimation, data of n consecutive moments are collected firstly, m time delays corresponding to the maximum m peaks on a cross power spectrum density distribution function at each momentare extracted, any time delay extracted from the m time delays is extracted for n consecutive moments, a connection line with the minimum time delay fluctuation is solved from n time delay connectionlines by using the dynamic programming algorithm, and besides, the peak of the cross power spectrum density distribution function is introduced to correct selection of a time delay connection line. After the time delay connection line with the minimum fluctuation is selected, the time delay at the last moment on the connection line is used as a signal TDOA, and by using a sliding time window, theabove process is repeated to solve a subsequent TDOA. Thus, the time delay coherence of multiple frames of mobile sound source signals are made full use of, the influences of multi-path effects on thesound source time delay are considered at the same time, the most correct time delay is extracted, and high value is achieved in the field of sound array processing.

Description

technical field [0001] The invention relates to the field of array signals, in particular to microphone array signal processing and arrival time delay estimation. Background technique [0002] The background of the present invention is based on actual needs. In recent years, when UAVs quickly became a research hotspot, it also brought a series of problems, such as UAV black flying, seriously affecting regional security. Therefore, drone defense is becoming a new area of ​​focus for governments and military forces. The sound of the drone has a distinctive signature that can effectively detect objects flying in the air. By setting up multiple microphone arrays, the signal of the UAV can be effectively collected, but the noise in the actual environment, the noise during the transmission of the signal in the circuit equipment, and some other reverberation, multipath interference, etc., the collected sound The signal is very noisy, which leads to a relatively large error in es...

AI Technical Summary

Problems solved by technology

By setting up multiple microphone arrays, the signal of the UAV can be effectively collected, but the noise in the actual environment, the noise during the transmission of the signal in the circuit equipment, and some other reverberation, multipath interference, etc., the collected sound The signal is very noisy, resulting in a relatively large error in estimating the time delay of signal arrival at the microphone (TDOA) when positioning the drone. In order to improve the accuracy of TDOA estimation, it can be improved when estimating TDOA

[0003] In the current research content, the TDOA estimate of the current moment is generally used to correct the TDOA estimate of the current moment, but this method only relies on the TDOA of the previous moment, which has poor redundancy and is prone to errors

Images