关键词: |
Obstacle avoidance, Navigation, Robots, Lidar, Global positioning systems, Ground vehicles, Autonomous vehicles, Operating systems, Algorithms, Laser rangefinding, Ros (robot operating system), Unstructured outdoor environments, Mobile robot navigation, Gps (global positioning systems), Matlab, Agv (autonomous ground vehicles), Obstacle avoidance algorithm, Lrf (laser rangefinders) |
摘要: |
The ability to detect, characterize, and avoid obstacles is a critical requirement for autonomous robotic systems, especially in dynamic environments. While autonomous vehicle research and development continues at a rapid pace, these systems are becoming more complex and expensive. The objective of this thesis was to determine the feasibility of utilizing a single two-dimensional laser scanning rangefinder for robust obstacle avoidance in unstructured outdoor environments. Specifically, sensing and control algorithms were developed for an autonomous ground vehicle (AGV). The system was designed to operate in varying outdoor environments while avoiding both static and dynamic obstacles. The AGV was able to effectively identify and avoid obstacles within its field of view and to navigate to specific coordinates across variable terrain. While this solution was limited by the sensor used and was not effective in all environmentssuch as when obstacles encountered were too short to enter the scanners plane of view the algorithm developed was successful for visible objects. Small improvements, such as using a gimballed scanner or one that scans in three dimensions, would make this solution more robust for a wider range of environments. |