一种时变洋流场下AUV最优能耗路径规划方法

在时变洋流场环境下,洋流矢量增加了时间维度,在时间角度上可进一步利用洋流以节约自主水下机器人(AUV)能量消耗.此外,在该环境中无后效性不再成立,基于经典贪婪策略的路径规划算法不再适用.鉴于此,结合路径参数选择和双层规划算法,提出一种适用于时变洋流场环境的能耗最优路径规划算法.出发时间和AUV推进速度均可以在时间维度上等待有利洋流,且推进速度与其能量消耗直接相关,因此,引入出发时间和推进速度作为路径参数.在此基础上,针对无后效性不成立问题,使用双层规划作为路径规划算法,分析该算法在时变洋流场环境下的适用性.算法将路径规划任务分为路径规划与路径优化两部分,路径规划部分采用蚁群系统算法构建通道,路径优化部分由量子粒子群算法对路径参数进一步优化,在保证全局最优的同时能够解决传统基于栅格的路径规划算法中机器人运动方向受限的问题.最后以Kongsberg/Hydroid REMUS 600s型水下机器人为模型,对所提出的路径规划算法进行仿真验证.

Energy-optimal path planning for AUV with time-variable ocean currents

Under the time-varying ocean current environment, the ocean current vector appends the time dimension, and the ocean currents can be further utilized to save autonomous underwater vehicles(AUVs) energy consumption in the temporal sense. In addition, classical greedy-based path planning algorithms are not applicable because non-aftereffect no longer holds in this environment. For the above reasons, an energy-optimal path planning algorithm in the time-varying ocean current environment is proposed, which combines the selection of path parameters and bilevel optimization. Firstly, both departure time and AUV propulsion velocity can wait for favorable ocean currents in time dimension, and AUV propulsion velocity is directly related to its energy consumption. So the departure time and propulsion velocity are introduced as path parameters. On this basis, the bilevel optimization is used as a path planning algorithm to solve the problem of non-aftereffect, and the applicability is analyzed. In the proposed approach, the task of path planning is divided into two parts: path planning and path optimization. In the path planning part, the ant colony system algorithm is used to construct the passageway, then the quantum particle swarm optimization algorithm is applied to further optimize the path parameters in the passageway at the path optimization part. The proposed algorithm ensures the global optimum of the resulting path and solves the problem of discrete motion directions caused by grid-based environment. Finally, to verify the validity of proposed scheme, several simulations, which the Kongsberg/Hydroid REMUS 600s is used as the model, are executed.