Robust, Adaptive and Dynamic Robotic Motion Planning

研究将在机器人运动规划的几个关键领域进行。重点研究自主机器人的运动规划算法。一个主要的焦点是研究不完全信息或环境信息变化的运动规划问题，这要求运动规划算法具有自适应能力。此外，该研究考虑了传感和控制中固有误差的运动问题，因此运动规划算法必须对缺乏机器人位置信息的情况具有鲁棒性，并且在可能的情况下，算法必须最小：即使用尽可能少的关于机器人位置和环境的信息。调查还将考虑动态变化环境中的运动规划问题；这将需要动态算法来响应不断变化的输入。这项研究将涉及使用尚未被充分研究的新配方解决运动规划问题。研究将特别集中在机器人运动规划的新兴领域，这些领域对自适应、鲁棒、最小和动态机器人算法的总体目标影响最大。要研究的主题包括针对随机和/或可能动态学习环境概率模型的未知环境的在线规划算法，以及追踪移动游戏（涉及与对手的移动规划）。

Research is to be conducted in several key areas of robotic movement planning. The emphasis is on robot motion planning algorithms for autonomous robots. A principal focus is on the investigation of motion planning problems with incomplete information or changing information about the environment, requiring the motion planning algorithm to be adaptive. Moreover, the research considers movement problems with inherent errors in the sensing and control, so the motion planning algorithms must be robust to a lack of information on the position of the robot, and if possible the algorithms must be minimal: that is use the least possible amount of information about the robot position and environment. The investigations will also consider motion planning problems in environments which are dynamically changing; these will require dynamic algorithms that respond to changing inputs. The research will involve solving motion planning problems using new formulations that have not yet been fully investigated. The research will concentrate particularly in emerging new areas in robotic movement planning that impact most strongly the overall goals of adaptive, robust, minimal, and dynamic robotic algorithms. Examples of topics to be researched are on-line planning algorithms for unknown environments which are randomized and/or may dynamically learn probabilistic models of the environment, and pursuit movement games (which involve movement planning with an adversary).