改进RRT*的室内机器人路径规划算法

Improved Indoor Robot Path Planning Algorithm of RRT*

摘要:

路径规划在室内机器人的应用中有着无可比拟的作用。为了提高路径规划算法收敛的速度，综合时间消耗和路径质量方面考虑，针对RRT*（渐进最优快速扩展随机树）算法的局限性，提出一种改进的RRT*算法。该算法采用目的性的设置采样点代替原算法中的随机高斯采样和引进人工势场与避障策略结合的思想，设置目标偏向性，引导随机树生长方向，然后利用曼哈顿距离代替欧几里得距离作为代价估值函数，防止陷入极小值以及一定程度上减小算法时间损耗。实验表明，该方法可有效平衡算法收敛时间与最佳路径的可靠性。

Abstract:

path planning plays an incomparable role in the application of indoor robot. In order to improve the convergence speed of path planning algorithm, and considering the time consumption and path quality, an improved RRT*( Rapidly Exploring Random Tree Star ) algorithm is proposed, aiming at the limitations of the algorithm. The algorithm adopts the purpose of setting instead of the original algorithm of random sampling gaussian sampling and the introduction of artificial potential field and obstacle avoidance strategy of combination of ideas, set up the target bias, guide the random tree growth direction, and then use the Manhattan distance to replace the Euclidean distance as the price valuation function, prevent a minimum loss of time and to a certain extent reduce algorithm. Experiments show that this method can effectively balance the convergence time and the reliability of the optimal path.