Warning:
This page is
provided for historical and archival purposes
only. While the seminar dates are correct, we offer no
guarantee of informational accuracy or link
validity. Contact information for the speakers, hosts and
seminar committee are certainly out of date.
The Robotics Institute Carnegie Mellon University
The task of planning trajectories for a mobile robot has received considerable attention in the research literature. Algorithms exist for handling a variety of robot shapes, configurations, motion constraints, and environments. Most of the work assumes the robot has a complete and accurate model of its environment before it begins to move; less attention has been paid to the problem of unknown or partially-known environments. This situation occurs for an exploratory robot or one that must move to a goal location without the benefit of a floorplan or terrain map. Existing approaches plan an initial global path based on known information and then modify the plan locally as the robot discovers obstacles with its sensors. This strategy works well in environments with small, sparse obstacles, but it can lead to grossly suboptimal and incomplete results in cluttered spaces. An alternative approach is to replan a global path from scratch each time a new obstacle is discovered. This brute-force approach is optimal, but it is grossly inefficient. This talk describes a new algorithm, D*, capable of planning paths in unknown, partially-known, and changing environments in an efficient, optimal, and complete manner. D* models the environment as a graph, where each node represents a robot state, and each arc represents the cost of moving between two states. Initially, a path is planned from the goal to the robot's location using known information. As the robot moves, its sensors discover obstacles in its path. These discoveries are handled by modifying the arc costs. D* propagates cost changes minimally to compute a new optimal path. The process repeats until the robot reaches the goal or determines that it cannot.
D* was implemented and compared head-to-head with the brute force replanner to verify optimality and determine the performance improvement. It was integrated with an obstacle avoidance system and tested on the Navlab II. The robot succeeded in finding a specified goal in a cluttered environment given no initial map information. The algorithm and the results will be presented.
Host: Yangsheng Xu (xu+@cs.cmu.edu) Appointment: Ava Cruse (avac@cs.cmu.edu)