基于USVSim的无人艇综合仿真平台设计与验证

Design and verification of unmanned surface vehicle integrated simulation platform based on USVSim

针对无人艇开发周期长、控制及感知模块难以综合测试等问题,提出了一种基于USVSim和MATLAB的无人艇综合仿真系统,实现无人艇航行控制系统快速开发和模拟测试的任务需求。该仿真系统由基于Matlab/Simulink的指令下发界面、基于Robot Operating System架构的无人艇控制系统和基于USVSim的虚拟仿真场景3部分组成。该系统为水面无人艇开发与测试提供了三维可视化仿真测试工具。该平台引入虚拟物理仿真引擎,使仿真结果更加直观。同时相机、激光雷达等常用的环境感知传感器可以作为插件应用在仿真器中,使仿真阶段贴合实际。平台具有开源的雷达、相机、GPS、姿态传感器、里程计等常用传感器,具有良好的可扩展性;也能根据平台硬件性能模拟不同海况或者多艇协同。以无人艇视线法制导算法为例,控制对象为自行研发1.75 m无人艇仿真模型,进行综合仿真性能测试。结果表明,该仿真系统达到预期效果,为后续无人艇感知-规划-控制的深入研究提供了验证平台。

To address the issues of difficult development and testing of unmanned surface vehicle and the difficulty of simulating the perception module,this paper proposed an integrated simulation system for unmanned surface vehicle based on USVSim(Unmanned Surface Vehicle Simulator)and MATLAB to achieve the task requirements of rapid development and easy learning and use of unmanned surface vehicle navigation control algorithms.The simulation system consists of three parts:the command issuing interface based on Matlab/Simulink,the unmanned surface vehicle control system based on Robot Operating System architecture and the virtual simulation scenario based on USVSim.The system provides an efficient simulation and testing tool for the development and testing of unmanned surface vehicle in 3D visualization.This platform introduced a virtual physical simulation engine to make the simulation results more intuitive.Meanwhile cameras,LIDAR and other common environment sensing sensors can be simulated in the simulator as plug-ins to fit the real situation in the simulation stage.Platform has open source radar,camera,GPS,attitude sensor,odometer and other common sensors,with good scalability;It can also simulate different sea conditions or multi-boat cooperation according to the platform hardware performance.In this paper,the line of sight method guidance algorithm for unmanned boats was used as an example,and the control object was a self-designed 1.75 m unmanned surface vehicle’s simulation model for simulation performance testing.Results show that the simulation system can achieve the expected results and provide a good verification platform for the subsequent research of perception-planning-control of unmanned surface vehicle.