Dynamic Perception And Obstacle Avoidance System Of A Marine Surface Robot

As a kind of marine robots,the marine surface robot integrates various technologies such as mechanical design,artificial intelligence and motion control,which has a great application values in the field of military and civilian.In this context,a marine surface robot experimental platform based on the catamaran model is developed in this paper,and explores dynamic perception and obstacle avoidance in this platform.Firstly,the overall structure of the perception and obstacle avoidance system for marine surface robot is designed.Through analysis and model selection,a buoy-type dual-machine dual-propeller marine surface robot system is developed.The whole system is divided into three parts:machine robot system,host computer system and emergency remote control system.Furthermore,the hardware structure design scheme is expounded,and the navigation,guidance and control algorithm architecture of the marine surface robot is proposed.Secondly,the basic experimental platform of marine surface robot perception and obstacle avoidance system is designed in this paper.Aiming at the hardware design requirements of the whole system and adopting the modular idea,the basic experimental platform of the marine surface robot obstacle avoidance system is built based on STM32 processor.In terms of software design,the program is hierarchically divided and then develops in a top-down software system.The construction of the experimental platform for marine surface robot perception and obstacle avoidance system lays the foundation for subsequent theoretical research and sea trial experiments.Thirdly,after completion of the experimental platform construction,the marine surface robot obstacle avoidance is researched in this paper.In the presence unforeseen environments,a dynamics-constrained global-local(DGL)hybrid collision avoidance scheme incorporating global path planning and local hierarchical architecture is created for the marine surface robot with constrained dynamics.By incorporating the marine surface robot safety area into Theta*-based searching,global path planning algorithm is proposed to optimally generate sufficiently sparse waypoints.Considering the problems of execution efficiency,energy efficiency,control accuracy,etc.,this paper adopts Line-of-sight(LOS)guidance strategy to complete the path following of the robot.In order to realize the avoidance of static and dynamic obstacles in the navigation process,a local hierarchy is established by fuzzy decision making(FDM)and fine dynamic window(FDW)layers,which are responsible for large-and close-range collision avoidance from unforeseen static and/or dynamic obstacles,respectively.By inserting virtual waypoints onto the globally optimal path,a seamless interface between global and local obstacle avoidance mechanism is devised,and thereby contributing to the entire DGL hybrid obstacle avoidance scheme.In terms of motion control,an anti-integral saturation PID heading controller is designed to improve the steering efficiency and stability of the robot.Finally,in order to verify the stability of the experimental platform and the effectiveness of the proposed algorithm,simulation analysis and sea trial experiments are carried out in this paper.The waypoint guidance,the tracking of the trailing path and the autonomous returning experiments demonstrate the effectiveness and superiority of the guidance control system.The laser radar ground-based experiment lays the hardware foundation for the realization of the dynamic perception and obstacle avoidance system.The simulation studies and comprehensive comparative analysis of obstacle avoidance algorithm in real scene map verify the effectiveness and superiority of the global-local hybrid obstacle avoidance strategy.