Control of a Hierarchical Team of Robots for Urban Search and Rescue

Abstract

Research teams worldwide are researching the application of robots for Urban Search and Rescue (USAR) operations and some are using teams of robots. The Mechatronics Research Group of Victoria University of Wellington is developing a low cost architecture of a team of USAR robots that is hierarchically structured and can operate autonomously. The objective of this thesis is to design the autonomous control system for the proposed architecture. The overall system design and combination of hardware and software solutions needs to be evaluated in a realistic environment. The project could not perform tests in a real environment and developed a realistic simulation environment instead to allow the evaluation of hardware and software constraints. This project successfully developed an incremental mapping algorithm which served as foundation for distributed path planning, and modified an existing navigation approach to cope with the main challenges of 3D operation environments. In order to deal with multiple robots, this thesis applied a centralised control mechanism and a combination of a global and local exploration strategy. This thesis contributes software solutions to operate the low cost robot architecture and identified weaknesses in the design of the middle tier of robots. The individual algorithms, and their combination in a major control system proved to be effective, but not without limitations. Consequently, this thesis suggests solutions to overcome some of these limitations.