Virtual road networks for long-term visual navigation of unmanned ground vehicles

点击翻译按钮获取中文摘要

Mobile robotics technology has advanced at a rapid pace in the last decade, making applications such as automated mining, agriculture, safety & security, and self-driving cars a near-term reality. The University of Toronto (UofT) has been developing a visual route-following technique for Unmanned Ground Vehicles (UGVs) that will serve as a key building block to enable this vision. To date, the technique has been tested primarily on point-to-point paths over short periods of time (i.e., a few days).The main goal of this proposal is to extend this core technique to allow an entire virtual road network (e.g., tens of kilometres of interconnected paths) to be defined and used for visual navigation of UGVs in a variety of service applications over long periods of time (i.e., months or years). This will require major advances in several areas including map maintenance, robust localization, and dynamic route planning. The project will be carried out in partnership with Clearpath Robotics Inc., an emerging world-class provider of unmanned vehicles located in Canada.This leap forwards in navigation technology will allow UofT to make core scientific advances in the area of long-term autonomy for mobile robots and will allow Clearpath to incorporate these advances in their future products and solutions, thereby expanding their customer base.