水下搜救中基于先验信息的AUV区域覆盖路径规划

Prior Information-based Coverage Path Planner for AUV Large-scale Search and Rescue

摘要:

搭载侧扫声呐(side-scan sonar，SSS)的自主水下航行器(autonomous underwater vehicle，AUV) 执行大规模海上搜救(search and rescue，SAR)任务时通常采用区域覆盖路径规划(coverage path planning， CPP)技术。由于任务失败的可能性很大，因此实现完全覆盖的同时，优先搜索目标可能存在的区域并提高侧扫声呐的数据质量十分必要。针对以上问题，提出了一种面向操作的全覆盖路径规划方法 SAS-A*，通过生成全覆盖路径，有效提高目标可能存在区域的覆盖速度和声呐数据的质量。该方法基于救援专家预测的目标位置和轨迹，采用二维高斯分布建立目标存在的概率模型，作为路径规划的先验信息。提出的 SAS-A*算法将下一个路径点选择策略转化为多目标决策问题，采用加权度量法来进行多目标优化，使 AUV 以更短路程、更少的转弯优先搜索目标可能出现的区域。仿真结果表明：SAR- A*路径规划方法适用于大规模搜救任务，可快速提高目标的搜索概率。

Abstract:

Coverage path planning(CPP)is usually applied to large-scale maritime search and rescue(SAR) mission conducted with an autonomous underwater vehicle(AUV)equipped with a side-scan sonar(SSS). There is a large chance of failure while AUV performs a large-scale mission. As a result，apart from completely covering the area，search of the area where the target might be preferentially and collecting of high-quality sensor data are needed. Aiming at this gap，a novel operation-oriented path planner named SAR-A* is proposed to generate a path for improving the coverage efficiency and quality of data obtained. Considering the predicted location and trace by rescue specialist，a two-dimensional Gaussian distribution is introduced to describe the probability map of target presence served as prior information. Moreover，the SAR-A* is developed for selecting the next waypoint from candidate cells. To choose the optimal next waypoint，the next waypoint selection strategy is then transformed into a multi-objective decision-making problem and the weighted metric method is applied to solve the multi-objective optimization. Due to the proposed SAR-A* algorithm，the AUV is guided to the most valuable area with a shorter path that has fewer turn，Simulation results show that the SAR-A* path planner is suitable for large-scale SAR，and can steeply increase confidence in locating the target.