Design And Implementation Of Sound Source Localization System Based On Time Delay Estimation

Positioning the sound source can be an effective solution to sound field enhancing, non-visible targeting and other issues. Currently a number of sound source localization algorithms have been studied, but there is a lot of work left before they can be used outside of laboratories. The main issue is that the algorithm has a highly complex operation which make the implementation of sound source localization system was difficult to achieve, so there is an urgent need to design and implement a low cost, high noise immunity, high performance and suitable real-time sound source localization algorithm.This paper has two aspects: one is the time delay estimation (TDE) based on adaptive filtering algorithm researching (referred to as: adaptive time delay estimation algorithm); another is design and implementation of sound source localization system based on the time delay estimation.The key-point of sound source localization system based on TDE is the TDE value, the mainly TDE algorithm has many problems such like: the complex calculation in algorithms, low performance at anti-jamming, low performance at system tracking ability. To make adaptive TDE algorithm has strong anti-interference, tracking ability, and easy to implement at embedded system, this paper has studied the TDE algorithm, by using the recursive least squares (RLS) adaptive TDE algorithm instead of least mean square (LMS) adaptive TDE algorithm. This paper also presents an improved variable step RLS adaptive TDE algorithm, the core factor is: by using the function to connect the step and the value of system error, to smooth the weights of system error in the adaptive algorithm. By doing so the final convergence rate of system is accelerated and the distortion in the steady state of the system is reduced. From the simulation results, compare to the LMS adaptive TDE algorithm, the variable step RLS adaptive TDE algorithm is not only inherited the highly anti-noise performance of the RLS adaptive TDE algorithm, but also improving the system tracking performance, speed up the operation speed, and create the conditions to achieve real-time sound source localization system.Based on the above principles, this paper design and implementation a sound source positioning system. The system hardware consists of a audio collection, sampling systems with three microphones, implementation of sound source location algorithm software is used the C language. Experimental indoor environment, sound source location system based on variable step RLS and LMS adaptive TDE algorithm, under the condition of the same intensity of the sound source signal and quiet environment for the location. The experiments have two parts: different sound sources Positions without the background noise, and the same position in different background noise. The results show that: under the condition of the same intensity of the sound source signal, the system positioning accuracy has decreased while distance from the sound source to the microphone increased. Sound source localization based on variable step RLS adaptive TDE algorithm is better than based on LMS adaptive TDE algorithm; under the condition of the same intensity of the sound source signal, by simulating the different intensity of environmental noise, experiments were carried out for the accuracy of the positioning system, while the noise intensity increases, the system positioning accuracy goes down. The accuracy of sound source localization based on variable step RLS adaptive TDE algorithm decreased under 65 decibels noise(GB UDC534.836), and accuracy began to decline, under 70 decibels noise environments. The accuracy of sound source localization based on LMS adaptive TDE algorithm decreased under 50 decibels noise, and accuracy began to decline, under 60,65,70 decibels noise environments, True potentiometer increased sharply.From the experimental results, the system has not only has a highly real-time location and tracking computing capability, but also with reasonable hardware development costs, this paper of sound source localization system has great value space at marketing.