CoverLib: Classifiers-equipped Experience Library by Iterative Problem Distribution Coverage Maximization for Domain-tuned Motion Planning

本文介绍了CoverLib，一种构建和利用经验库的原则性方法，用于快速运动规划。CoverLib通过迭代地向库中添加经验-分类器对来构建经验库，其中每个分类器对应于问题空间内可适应的经验区域。这个迭代过程是一个主动的过程，因为它选择下一个经验是基于其能够有效地覆盖未覆盖区域的能力。在查询阶段，这些分类器被用于选择一个预计可适应于给定问题的经验。实验结果表明，CoverLib有效地缓解了全局（例如基于采样的）和局部（例如基于优化的）方法之间的可规划性和速度之间的权衡。因此，它在问题域上实现了快速规划和高成功率。此外，由于其适应算法不可知的特性，CoverLib可以与各种适应方法（包括基于非线性规划和基于采样的算法）无缝集成。

CoverLib: A Cover-Based Approach to Library-Based Motion Planning

CoverLib is a principled approach for constructing and utilizing a library for fast motion planning. It iteratively adds an experience-classifier-pair to the library, where each classifier corresponds to an adaptable region of the experience within the problem space. During the query phase, these classifiers are utilized to select an experience that is expected to be adaptable for a given problem.

CoverLib effectively mitigates the trade-off between plannability and speed observed in global and local methods, achieving both fast planning and high success rates over the problem domain. It seamlessly integrates with various adaptation methods and is adaptation-algorithm-agnostic. Experimental results demonstrate the effectiveness of CoverLib.

Related work includes sampling-based and optimization-based motion planning methods, as well as other library-based approaches such as Goal Babbling and Experience Graphs.