Research And Implementation Of Multi-beam 3D Collision Avoidance Algorithm Based On C6657

With the depletion of land resources,the development of marine resources has gradually attracted people’s attention,and people have begun to use underwater robots to carry out some difficult underwater tasks.In order to ensure the successful completion of the UUV’s work,it is necessary to perform real-time obstacle recognition of the UUV in the complex underwater environment to solve the safety problems that it may encounter during the navigation process,so sonar is needed.Conventional forward looking sonar can only obtain the distance and azimuth parameters between the target and the carrier.This sonar system has obvious shortcomings,such as the lack of pitch angle estimation capability,which may seriously threaten the safety of carrier navigation.In order to solve the shortcomings of the traditional sonar system,this paper carries out the research of multi-beam 3D collision avoidance algorithm based on the 2D sparse array and realizes it on DSP,focusing on solving the problems of underwater target distance,azimuth and elevation angle estimation.Based on theoretical analysis,the thesis conducts a Matlab simulation verification study on the multi-beam three-dimensional collision avoidance algorithm.For underwater target distance estimation,the energy value corresponding to each beam is obtained by the energy center detection algorithm based on the beam formed data.After a series of thresholds,the echo arrival time is obtained,and then the target obstacle and carrier are obtained.Use the minimum angle theorem and the phase difference relationship to obtain the horizontal and vertical angles of the target obstacle;finally combine the ground coordinate position between the target obstacle and the platform where the detection sonar is located,roll,pitch,and ship direction The angle and other information are converted into Gaussian coordinates and uploaded to the host computer.The paper also carried out the real-time realization research of the algorithm on the multi-core DSPC6657,taking into account the real-time performance,and further verifying and optimizing the algorithm,realizing the real-time detection of target obstacles based on the multi-beam sonar image.The sonar prototype has undergone many field experiments in the field,including pond experiment,Songhua Lake experiment,and Shuibuya experiment,verifying the feasibility of the signal processing scheme and the real-time nature of the algorithm.