| During the past decade, research in mobile robotics has transcended the indoor navigation stage. Expanding the application to outdoor navigation, the research community made significant strides recently in terms of vehicle autonomy in DARPA Urban Challenge. However, these successes heavily relied on inertial/GPS navigation systems costing up to ;Our vision includes smart wheelchair systems eventually capable of autonomous navigation in unstructured, outdoor environments. As our primary work in this area, the Automated Transport and Retrieval System (ATRS) enables independent mobility for drivers in wheelchairs by automating the stowing and retrieval of the driver's wheelchair system. While ATRS has been commercialized, and its smart wheelchair system does in fact navigate autonomously, its autonomy is limited to an area immediately adjacent to the host vehicle. We would like to build on these results to support a greater range of smart wheelchair applications. Key to this objective is a robust means for outdoor localization.;This dissertation proposes a map-based localization approach. In this approach, three-dimensional map data are acquired by vehicles employing high precision inertial/GPS systems for pose estimation, in conjunction with high-fidelity light detection and ranging (LIDAR) systems whose range measurements are subsequently registered to a global coordinate frame. The resulting map data are then synthesized a priori to identify robust, salient features for use as landmarks in localization. By leveraging such maps with landmark meta-data, robots possessing far lower cost sensor suites gain many of the benefits obtained from the higher fidelity sensors, but without the cost. Experiments show that by using such a map-based localization approach, a smart wheelchair testing platform outfitted only with a 2-D LIDAR and encoders was able to maintain accurate, global pose estimates outdoors over paths of almost 1km in absence of GPS. |
