基于电子海图矢量数据建模的无人船路径规划

Path Planning of Autonomous Ship Based on Electronic Chart Vector Data Modeling

栅格地图在自主船路径规划领域应用广泛。为解决栅格地图在表示多种类型障碍物共存的复杂环境时所存在精度不高及路径可靠性不足的问题,提出一种基于电子海图矢量数据——点、线、面要素建立精确环境势场的方法,可用于描述不同类型障碍物及狭窄水域。在此基础上,改进了基于静态环境势场及动态环境引力和斥力的混合路径规划算法,并结合Mariner典型船型及比例微分控制,进行了多船多障碍动态环境和狭窄水域静态环境的仿真测试。试验结果表明,该算法实现了《国际海上避碰规则》约束下的船舶实时自主避碰,并兼顾多种形状和类型的障碍物。建模方法简单实用,具有与电子海图显示与信息系统相结合的潜力。

Image grid maps are widely used in path planning for autonomous ships. However, when representing complex environment where several types of obstacles coexist, its inaccurate accuracy and insufficient path reliability are problems. In order to solve the problems, a new method of potential field based on vector data of electronic chart, such as points, lines, and surfaces, is proposed to represent accurate environment which may including different types of obstacles and narrow waters. On this basis, a hybrid path-planning algorithm is also improved through calculating the minimum potential value of the static environment and combining attractive and repulsive forces of potential field for dynamic environment. Then, this model is applied to Mariner type ships and proportional & differential (PD) controller, to conduct simulations for the static environment of narrow waters and for the dynamic environment with multi-ship and multi-obstacle. Simulation results show that this algorithm can cope with the problems of real-time collision avoidance of autonomous ships under constraints of the International Regulations for Preventing Collisions at Sea, and takes various shapes and types of obstacles into account. The modeling method is simple and practical, and has potential to be combined with Electronic Chart Display Information Systems.