Motion Capture System and Mobile Robot Vehicle for Indoor Autonomous Navigation Research

点击翻译按钮获取中文摘要

A central issue in robotics research is to determine the position of an entity of interest accurately in the physical space such as the position of a robot, a visual landmark or an object being manipulated by a robot. This information is critical for quantitatively evaluating the performance of a robot localization and mapping algorithm or method, for creating benchmarks and ground truth to compare competing robotics solutions, and sometimes for sensing the robot position for its control. Although originally developed for capturing human movements, motion capture has become a widely used measurement tool in robotics to meet the needs described above. This proposal requests funding for such a system in order to allow robotics researchers at the University of Alberta to conduct their studies effectively. The PI and Co-PI of this proposal actively conduct research in autonomous robot navigation and mobile robot manipulation, in the Department of Computing Science. The PI’s work in simultaneous localization and mapping or SLAM attempts to build an accurate three-dimensional description of the environment (map) while localizing the robot within the map at the same time. The Co-PI, Professor Martin Jagersand, is interested in object manipulation by a robot arm on a mobile platform where a vision system must determine the position of the object sufficiently accurately for its manipulation. All these studies critically depend on the availability of the requested motion capture system, in order to physically implement robot navigation or control algorithms and quantify their performance. The requested motion capture system represents a shared infrastructure that places the U of A on par with leading robotics research institutions in the world. Finally, a ground robot is also requested to replace the aging indoor mobile robots (a Pioneer robot and a NAO robot) being used in the Robotics Research Laboratory to study autonomous robot navigation.