Research And Implementation Of Autonomous Obstacle Avoidance Method For Unmanned Underwater Vehicle Based On Forward Looking Sonar

Building a maritime power which requires the rapid development of various high-tech and equipment including unmanned underwater vehicles.As one of the important technologies of UUV,autonomous obstacle avoidance technology is a current research hotspot.This paper studies the autonomous obstacle avoidance method of UUV based on forward-looking sonar,and divides autonomous obstacle avoidance into two parts: obstacle detection and path planning.Obstacle detection: According to the image characteristics of two-dimensional forwardlooking sonar,through the analysis and comparison of classic image filtering,image enhancement and image segmentation algorithms,bilateral filtering,piecewise linear transformation,and K-means segmentation are used as an important part of the ostacle detection algorithm.Finally,the connected area is analyzed by the two-pass scanning method to obtain the basic information of the obstacle.Firstly,according to the image characteristics of 2D forward-looking sonar,the classical image filtering,image enhancement and image segmentation algorithms are analyzed and compared through subjective vision and objective indicators,the image segmentation method combining bilateral filtering,piecewise linear transformation,Otsu threshold segmentation and K-means clustering is used as the detection method of forward-looking sonar obstacles,Finally,the two-pass scanning method is used to analyze the connected area,and the basic information of the obstacle is calculated.Then,according to the mastery of the information,the path planning of the UUV is divided into global and local path planning.Through the simulation description of typical global and local path planning methods and the analysis of the properties of the algorithms,it is considered that the A* algorithm and the BUG2 algorithm Global and local path planning for unmanned underwater vehicles.The A* algorithm is improved through environmental modeling,selection of appropriate heuristic functions,and path optimization.The improved A* algorithm has faster search speed,shorter paths,and improved security.The BUG2 algorithm is improved through obstacle processing,selection of surrounding directions,and path generation.The improved BUG2 algorithm has a shorter planning time and improved safety performance.Finally,the autonomous obstacle avoidance software of UUV is designed based on QT,and the navigation process of UUV is simulated by using the data collected from Songhua Lake,the autonomous obstacle avoidance software is verified,and the obstacles are also proved.The feasibility of the detection method,improved A* algorithm,and improved BUG2 algorithm proves the feasibility of the autonomous obstacle avoidance method for UUVs studied in this paper.