Acoustic Target Location Algorithm Based on Zigbee Wireless Sensor Network.docVIP

Acoustic Target Location Algorithm Based on Zigbee Wireless Sensor Network 　　Abstract. The intensity of the sound energy received by the wireless sensor networks ，using independent component analysis method to obtain the goal of energy between each other mixed proportion， and the use of sound energy model and maximum likelihood estimation method to locate the target. Experimental results show that the method has a higher positioning accuracy and noise immunity. 　　Key words： wireless sensor networks； Zigbee； maximum likelihood estimation； targeting 　　Introduction 　　Positioning of wireless sensor networks has become a hot spot in wireless sensor network research recent years. Currently， for the positioning of the sound target， there are three main method： time difference of arrival （TDOA）， based on the positioning of the direction of arrival （DOA， direction of arrival） and based on the received signal energy positioning . Requirement of TDOA measurement for time difference accuracy is high， only for a single target location； DOA for narrowband signal. Energy positioning method is simple， that can be used for the positioning of multiple targets， effectively saving the energy of the sensor nodes. 　　Currently， the targeting method based on the energy model is： CPA， maximum likelihood estimation method， the energy ratio method. CPA simple but can only be used for single target low positioning accuracy； energy ratio law generally solved using the least squares method and nonlinear least squares method， fast positioning， but affected by noise and can only locate a single target； high positioning accuracy of the estimation method of maximum likelihood， appear on multiple targets at the same time is also effective， energy-based multi-targeting method， but its positioning accuracy of less than a single target positioning accuracy. In this paper， the independent component analysis （ICA） obtained mixing matrix from mixed-signal， combined with the maximum likelihood esti