| Popis: |
Basic abilities for robots in rescue environments consisted of autonomous exploration and target recognition. In this essay, we combine these two abilities and propose a target-based robot autonomous exploration method. When no victim is found, robots will explore the rescue environments autonomously using the wall following and frontier-based techniques. Once finding a victim, robots will use the recognized victim as the navigation goal, plan a path use the "timed elastic band" approach, and quickly reach the victim to facilitate further rescuing. By combining the autonomous exploration and the victim recognition, our method can improve the efficiencies of both tasks. Extensive experiments and tests have been conducted on a real rescue robot. The results of experiments prove that our method can enable robots to explore the rescue environments autonomously with high efficiency. In the RoboCup Rescue Robot League (RRL) competitions, we have achieved good results by using this method. |
