泡泡一分钟：Automatic Parameter Tuning of Motion Planning Algorithms

Automatic Parameter Tuning of Motion Planning Algorithms

运动规划算法的自动参数整定

Jos´e Cano, Yiming Yang, Bruno Bodin, Vijay Nagarajan, and Michael O’Boyle

张宁 Automatic Parameter Tuning of Motion Planning Algorithms
 https://pan.baidu.com/s/17rNCxNp3Lqbtqt-sO1xhiw

张宁

Abstract—Motionplanningalgorithmsattempttofindagood compromise between planning time and quality of solution. Due to their heuristic nature, they are typically configured with several parameters. In this paper we demonstrate that, in many scenarios, the widely used default parameter values are not ideal. However, finding the best parameters to optimise some metric(s) is not trivial because the size of the parameter space can be large. We evaluate and compare the efficiency of four different methods (i.e. random sampling, AUC-Bandit, random forest,andbayesianoptimisation)totunetheparametersoftwo motion planning algorithms, BKPIECE and RRT-connect. We present a table-top-reaching scenario where the seven degreesof-freedom KUKA LWR robotic arm has to move from an initial to a goal pose in the presence of several objects in the environment. We show that the best methods for BKPIECE (AUC-Bandit) and RRT-Connect (random forest) improve the performance by 4.5x and 1.26x on average respectively. Then, we generate a set of random scenarios of increasing complexity, and we observe that optimal parameters found in simple environments perform well in more complex scenarios. Finally, we findthatthetimerequiredtoevaluateparameterconfigurations can be reduced by more than 2/3 with low error. Overall, our results demonstrate that for a variety of motion planning problems it is possible to find solutions that significantly improve the performance over default configurations while requiring very reasonable computation times.

运动规划算法试图在规划时间和解决方案质量之间做出妥协。由于它们具有启发性，它们通常具有多个参数。在本文中，我们证明了在许多情况下，广泛使用的默认参数值并不理想。但是，找到优化某些度量标准的最佳参数并非易事，因为参数空间的大小可能很大。我们评估和比较四种不同方法（即随机抽样，AUC-Bandit，随机森林和贝叶斯优化）的效率，以及两个运动规划算法，BKPIECE和RRT-connect的参数。我们提出了一种桌面式的场景，其中七个自由度的KUKA LWR机器人手臂必须在环境中存在多个物体的情况下从初始姿势移动到目标姿势。我们表明，BKPIECE（AUC-Bandit）和RRT-Connect（随机森林）的最佳方法分别平均提高了4.5x和1.26x的性能。然后，我们生成一组增加复杂度的随机场景，并且我们观察到在简单环境中找到的最佳参数在更复杂的场景中表现良好。最后，我们发现评估参数配置所需的时间可以减少超过2/3而且误差很小。总的来说，我们的结果表明，对于各种运动规划问题，可以找到在默认配置下显着提高性能同时需要非常合理的计算时间的解决方案。