Abstract

We present the motion planning framework for Boss, Carnegie Mellon’s winning entry in the Urban Grand Challenge. Urban environments present a number of motion planning challenges, including high-speed operation, complex inter-vehicle interaction, parking in large unstructured lots, and highly constrained maneuvers. Our approach combines a local planner that utilizes a model-based trajectory generation algorithm for computing dynamically feasible actions with two global planners for efficiently generating long range plans in both on-road and unstructured areas of the environment. This talk will include examples and illustrations from both testing and the Urban Challenge itself.

Speaker Biography

Dave Ferguson is a Research Scientist at Intel Research Pittsburgh, working on planning and coordination for single agents and multi-agent teams. He is currently co-PI of the Personal Robotics project, in which an anthropomorphic robotic arm and mobile robot base coordinate to accomplish useful manipulation tasks in populated indoor environments. He was also the planning lead for Carnegie Mellon’s Tartan Racing team, winners of the 2007 DARPA Urban Challenge autonomous vehicle race, and the initial software lead for Carnegie Mellon’s Mine Mapping Project, in which an autonomous robotic vehicle successfully mapped over 300 meters of an abandoned coal mine. In 2007, one of his planning algorithms was incorporated into the onboard navigation framework of the Mars Exploration Rovers. Dave obtained his Ph.D. from the Robotics Institute at Carnegie Mellon University in 2006.