Design And Implementation Of Autonomous Navigation Obstacle Avoidance System For Unmanned Surface Vehicle

With the rapid development of “the Belt and Road” strategy in China,“21st-Century Maritime Silk Road” has been highlighted by an increasing number of countries and regions.The water traffic is increasingly busy and ships begin to develop in the direction of high-speed and intelligence.In addition,the obstacle avoidance of autonomous navigation technology of unmanned surface vehicle is also widely concerned.Based on computer vision,this paper studies an obstacle avoidance of autonomous navigation system of unmanned surface vehicle,which can realize autonomous navigation and obstacle avoidance in lakes and rivers.The main content of this paper includes: the design and model selection of the hardware system of the unmanned surface vehicle,including power module,obstacle avoidance of autonomous navigation module and wireless communication module.Raspberry Pi is connected to the camera,which first compresses and encodes the video data collected by Open CV,and then transmits data to the host computer through the wireless bridge.The host computer decodes the received video data and obtains the water surface video.The host computer uses the EDlines-based linear detection algorithm to detect waterfront lines and water-sky lines,so as to segment the water surface image.Then it introduces the deep learning object detection method to realize the detection and identification of obstacles such as ships,buoys,reefs,etc.,and obtain the outline of obstacles based on the improved SSD algorithm of Mobile Net-V2.Meanwhile,it can detect the distance with the combination of microwave ranging radar sensor,so that the unmanned surface vehicle can bypass the obstacles and make it pass through safely.Through operating host computer software on the computer,the starting point,route points along the route and final target points of the unmanned surface vehicle can be set on the map.The navigation trajectory can be displayed on the map,and the surface of the water and obstacles can be displayed in real time.If necessary,the autonomous navigation obstacle avoidance operation mode can be switched to manual operation mode.Finally,the autonomous navigation obstacle avoidance system is verified through software testing and launching experiments.The experimental results verify the feasibility of the designed system.