RIA: Parallel Artificial Intelligence Techniques Applied toRobot Planning

点击翻译按钮获取中文摘要

The development of intelligent robot planning techniques is an essential element of future underwater and space exploration and industrial automation. These technologies find application in autonomous underwater and planetary rovers, autonomous rendezvous and docking, operation and maintenance of a nuclear power plant, extravehicular activity, autonomous landers, robotic and telerobotic control and automation, and in the general construction and operation of a laboratory, ship or industrial site. The purpose of this project is to increase the efficiency and effectiveness of robot planning by making use of parallel hardware and machine learning techniques. The project is made up to two component parts, each of which contributes independently and collectively to the automation of robotic activity. The first subproject focuses on developing parallel algorithms for heuristic search, which is a fundamental component in robot planning. The second subproject focuses on incremental modification and reuse of plans by searching for similarities between old and new plans. These two projects have been developed and tested on several search and planning problems. We propose to extend the existing research by improving the parallel search algorithms, developing parallel planning algorithms, and applying dynamic graph match to allow reuse of plans in a dynamically changing environment.//