Gaining diagnostic information for fault isolation

Robots in dynamic and uncertain environments are vulnerable to mission failures due to external perturbation or internal malfunctions. Diagnosis is the process to detect, locate or even assess the fault. Since robots rely on their function modules to sense the external environment, it is difficult to locate the fault under uncertainties of robot components. The situation can be worse when there is also uncertainty about the environment. To resolve this issue, this paper proposes a new method to actively gain diagnostic information to locate the failure-cause more accurately under uncertainties. An integrated strategy of self-function-checking and diagnostic-plan is described. Validation using JSHOP2 planner showed that robots using this strategy was able to locate failure-cause with high autonomy.