A mechanism for real-time decision making and system maintenance for resource constrained robotic systems through ReFrESH

Robot operating environments and the status of robots are complex and varying, so it is practically impossible for a robotics designer to anticipate all system configurations to successfully complete a task prior to deployment. Therefore, a mechanism for dynamic decision making and configuration synthesis that copes with system fault and uncertainty is necessary. This paper implements such a mechanism within a self-adaptive framework (ReFrESH). The goal of this presented mechanism is to provide diagnosability and maintainability to manage the system performance during task execution in the presence of unexpected uncertainties. Specifically, the functionality of the proposed mechanism include: (1) detection of system performance degradation; (2) diagnosis and locate of the fault module; (3) synthesis of feasible task configurations; (4) selection of the optimal one. We illustrate the feasibility of the proposed mechanism through a visual servoing task.