Task-based Hardware Reconfiguration in Mobile Robots Using FPGAs

This paper presents a methodology for the realization of intelligent, task-based reconfiguration of the computational hardware for mobile robot applications. Task requirements are first partitioned into requirements on the system hardware and software. Architecture is proposed that enables these requirements to be addressed through appropriate hardware and software components. Hardware–software co-design and hardware reconfiguration are utilized to design robotic systems that are fault-tolerant and have improved reliability. It is shown that this design enables the implementation of efficient controllers for each task of the robot thereby permitting better operational efficiency using fixed computational resources. The approach is validated through case studies where a team of robots is configured and the behavior of the robots is dynamically modified at run-time. It is demonstrated through this implementation that the design procedure results in increased flexibility in configuration at run-time. The ability to reconfigure the resources also aids collaboration between robots, and results in improved performance and fault tolerance.