An open architecture for on-board intelligent diagnosis and recovery in autonomous robots

A way to improve thc pcrformancc in autonomous robots is adding intclligcncc to control systcms avoiding hardware rcdundancy and inopportunc mission abortion. In this papcr, wc prcscnt a distributcd architccturc supporting diagnosis and control of autonotnous robots with thc purposc of assisting the operator or piloting systcm in managing fault detection, risk asscssmcnt, and rccovcry plans under unccrtainty. Thc architccturc is gcncdc, open and modular consisting of a decision modulc, intclligcnt modulcs, and robot piloting modulcs. Thc dccision modulc is thc architccturc corc rcqucsting intclligcnt modulcs for diagnosis and rccovcry action bascd on data obtaincd from robot piloting modulcs.