Development of an integrated mobile robot system at Carnegie Mellon University : June 1988 annual report

This report describes progress in development of an integrated mobile robot system at the Carnegie Mellon Robotics Institute from July 1987 to June 1988. This research was sponsored by the Defense Advanced Research Projects Agency and monitored by the US Army Engineer Topographic Laboratories under contract DACA76-86-C-0019. Our program includes a broad agenda of research in the development of mobile robot vehicles, focused on the NAVLABcomputer-controlled van. In the year covered by this report, we addressed major issues in both hardware and software for autonomous mobile robots: • Evolution of the NAVLAB Vehicle. We built the NAVLAB mobile robot vehicle in our previous work under this contract, by outfitting a commercial truck chassis with computercontrolled drive and steering controls and a set of on-board computer workstations. The NAVLAB serves as a mobile navigation laboratory that allows researchers to interact intensively with the system during testing and execution. This year has seen a continued evolution and improvement of the NAVLAB mechanism, sensors, controller, and Virtual Vehicle interface to higher-level planning and perception software. • Evolution of the CODGER Blackboard. Last year, as part of this research program, we designed and implemented the CODGER blackboard system for robot perception and reasoning on a distributed collection of processors. This year, in response to our experience in using CODGER for mobile robot control, we have upgraded it to deal with geometric models and uncertainty in perception and map data. • Experiments With the Driving Pipeline. To control the NAVLAB and Terregator mobile robot vehicles, we developed the Driving Pipeline architecture last year for coordinating road following, obstacle avoidance, and vehicle motion control. In our ongoing research, we have performed numerous experiments with this system that demonstrate its value. This hardware and software is the basis for the New Generation System (NGS) for robot vision and navigation, which integrates many independent technologies to produce an integrated mobile robot system.