基于改进的RRT^(*)算法的AUV集群路径规划

AUV cluster path planning based on improved RRT^(*)algorithm

[目的]针对微小型欠驱动自主式水下机器人(autonomous underwater vehicle,AUV)集群控制问题,设计一种基于改进RRT^(*)算法的编队控制策略。[方法]RRT^(*)算法规划的路径陡变难以跟踪且收敛速度较慢,针对该问题提出改进方法。首先加入偏置函数使随机采样点靠近目标点,然后采用Dubins曲线平滑连接采样点,通过在可变半径范围内重新布线,并设计有关曲线长度与避障的代价函数,选择最优路径。依据代价和最小值为多AUV分配集结点,协调多AUV速度完成最小集结时间约束,随后设计基于Dubins路径的分段向量场构造方法,使得多AUV跟踪规划路径,到达目标集结点时速度与方向保持一致。[结果]仿真结果表明,多AUV编队平均路径长度缩短26.6%,平均集结时间缩短21.7%。[结论]该算法路径规划质量高,可顺利完成编队集结任务。

[Objective]Aiming at the cluster control problem of small underactuated autonomous underwater vehicles(AUVs),a formation control strategy based on an improved RRT^(*)algorithm is designed.[Method]Paths planned by the RRT^(*)algorithm are steep and difficult to track,with slow convergence speed,so an improved method is proposed to solve the above problems.First,a deviation function is added to bring the random sampling points closer to the target point,then the sampling points are connected smoothly using a Dubins curve.By rerouting within the variable radius range and designing the cost function in relation to the curve length and obstacle avoidance,the best path is chosen.According to the cost and minimum value,multiple AUVs are assigned a rendezvous point,and the speed of multiple AUVs is coordinated to complete the minimum rendezvous time constraint.A segmented vector field construction method based on the Dubins path is then designed,enabling multiple AUVs to track the planned path and reach the target rendezvous point with the direction remaining the same.[Results]The simulation results show that the average path length of multiple AUV formations is shortened by 26.6%and the average assembly time is shortened by 21.7%.[Conclusion]The improved algorithm proposed herein has high path planning quality and can successfully complete formation assembly tasks.