大型自主水下机器人全局静态与局部动态融合路径规划方法

Research on fused path planning method of global static and local dynamic of large displacement autonomous underwater vehicles

针对大型自主水下机器人在做全局路径规划时面临环境建模复杂,算法求解能力弱以及面对局部动态障碍时自主性低,避障路径规划困难等问题,采用极坐标表示形成路径同心圆,在严格机动性约束下提出基于改进粒子群算法和速度障碍法的全局静态与局部动态相融合的路径规划方法。在极坐标表示的环境模型中,在全局静态规划中引入最优粒子"变异"过程提升算法求解能力;在局部动态规划中利用速度障碍法求解局部碰撞范围和安全路径区域以保证避障路径最优。实验结果表明,与传统粒子群和遗传算法相比,改进方法在全局静态规划中路径更短、求解能力更强,局部动态规划能够得到出最优避障路径。

When the large displacement autonomous underwater vehicle plans a path,there exist various problems,such as the difficulty in environmental modelling and the weak solution ability of algorithms facing global static environment,low autonomy and difficulties of path planning are exist in the process of obstacle avoidance planning in the face of local dynamic obstacles.To solve these problems,form concentric circles of path are used in polar coordinates,and a new path planning algorithm fusing global static and local dynamic which based on improved particle swarm optimization algorithm and velocity obstacle method is proposed under strict mobility constrains.In the environment model represented by polar coordinates,the optimal particle"mutation"process is introduced into the global static planning to enhance the algorithm’s solving ability;In the process of local dynamic planning,the velocity obstacle method is used to solve the local collision range and safe path area to acquire the optimal obstacle avoidance path.The results of simulate show that compared with the traditional particle swarm optimization and genetic algorithm,the improved method has shorter path and stronger solution ability in global static planning,and can also plan the optimal path in local dynamic planning.