Autonomous Coordination and Online Motion Modeling for Mobile Robots

摘要

Robots are rapidly becoming more involved in everyday military operations. As robots become more capable, their tasks will increase to include such roles as exploring enemy controlled buildings and caves. The goal of this thesis is to explore methodologies that allow robots to operate more autonomously. The first goal is to develop an algorithm that allows groups of robots to construct controlled formations with only local information. Experiments investigate the ability of this algorithm to handle obstacles, dynamic conditions, and varying number of robots. The second goal of this work is to demonstrate a method by which a robot can automatically determine how it is moving. Experiments demonstrate the ability of the algorithm to learn new models given models from other surfaces and robots. This work facilitates further research into creating complex formations using only local information and in fully automating current Simultaneous Localization And Mapping (SLAM) applications.