Seminars at the Faculty of Informatics

Talk@IDSIA: Markus Rickert - Balancing Exploration and Exploitation in Motion Planning

Monday, 25th of January, 12h00

Sala Primavera, Galleria 2, 6928 Manno

 

Computationally efficient robot motion planning must avoid exhaustive exploration of high-dimensional configuration spaces by leveraging the structure present in real-world planning problems. One way to accomplish this is by carefully balancing exploration and exploitation. Exploration seeks to understand configuration space, irrespective of the planning problem, while exploitation acts to solve the problem, given the available information obtained by exploration.
The presented Exploring/Exploiting Tree (EET) planner balances its exploration and exploitation behavior by acquiring workspace information and subsequently using this information for exploitation in configuration space. If exploitation fails in difficult regions, the planner gradually shifts its behavior towards exploration. Experimental results demonstrate that adaptive balancing of exploration and exploitation leads to significant performance improvements, compared to other state-of-the-art sampling-based planners.
A set of open source libraries have been developed as part of the reference implementation and various other robotic projects. Several basic data types, algorithms and abstractions are included, e.g., basic mathematics, kinematics, dynamics, hardware control, collision detection and motion planning.

Computationally efficient robot motion planning must avoid exhaustive exploration of high-dimensional configuration spaces by leveraging the structure present in real-world planning problems. One way to accomplish this is by carefully balancing exploration and exploitation. Exploration seeks to understand configuration space, irrespective of the planning problem, while exploitation acts to solve the problem, given the available information obtained by exploration.The presented Exploring/Exploiting Tree (EET) planner balances its exploration and exploitation behavior by acquiring workspace information and subsequently using this information for exploitation in configuration space. If exploitation fails in difficult regions, the planner gradually shifts its behavior towards exploration. Experimental results demonstrate that adaptive balancing of exploration and exploitation leads to significant performance improvements, compared to other state-of-the-art sampling-based planners.A set of open source libraries have been developed as part of the reference implementation and various other robotic projects. Several basic data types, algorithms and abstractions are included, e.g., basic mathematics, kinematics, dynamics, hardware control, collision detection and motion planning.