EPSCoR: Autonomous Fault-Tolerant Operation of Redundant Robotic Arms

Autonomous Fault-Tolerant Operation of Redundant Robot Arms Experiencing Locked Joint Failures Overview: The scienti?c objective of this project is to develop a comprehensive solution to realize au- tonomous fault-tolerant and fail-active operation for redundant robot arms experiencing joint failures. The central hypothesis of this project is that common electrical faults with the joint motors and drive electronics can be prognosed and diagnosed with online analytics of the fault signature information. The reliability and robustness of the robot system can then be optimized through fault-tolerant motion planning and kinematic design of optimally fault-tolerant robots. Three research goals are proposed: (1) Fault prognostics and diag- nostics of robot joints, including the joint motor faults and the faults with the drive electronics. (2) Compute the reliability map that represents the probability of task completion at different workspace locations for the predicted joint failure probabilities, including both positions and orientations. Then, this reliability map will be used to generate an optimal end-effector trajectory to maximize the probability of task completion. (3) Design optimal fault-tolerant robots to maximize the volume of the fault-tolerant workspace. The outcomes of this project will signi?cantly improve the reliability and robustness of the redundant robot systems that are operated in remote or hazardous environments, such as space exploration, underwater exploration, and nuclear waste remediation. The outcomes of this project can also be applied to safety-critical applications, such as human-robot interaction and robotic surgery. Intellectual Merit: We propose a fundamental scienti?c contribution for developing autonomous fault- tolerant and fail-active strategies for redundant robot arms based on predicted and identi?ed joint faults. In particular, the proposed prognostic and diagnostic algorithms differ from all the conventional approaches in that it does not require any additional sensors or hardware components, but still maintains an accurate and fast prognostic and diagnostic speed. In addition, most of the conventional fault-tolerant control methods focus only on failure recovery, and unfortunately it is usually too late to mitigate damages after failures happen. The proposed fault-tolerant motion planning and fault-tolerant design methods maximize the task completion probability in anticipation of all potential failures. Finally, existing fault-tolerant motion plan- ning algorithms only consider the local fault-tolerant performance or assume that all the joints are equally likely to fail. In this proposed research, both the global fault-tolerant performance and different joint failure probabilities will be considered in the motion planning and kinematic design. Broader Impacts: (1) This project will dramatically improve the reliability and fault-tolerant capability of redundant robot arms in robotics exploration missions or safety-critical applications. (2) The PI and Co-PI will devote signi?cant efforts to recruit women and underrepresented minority students in STEM ?elds to participate in this project. In particular, the PI and Co-PI will recruit black and female students in STEM from the local chapters of National Society of Black Engineers and Society of Women Engineers. The PI will also facilitate internship programs for undergraduate students from Kentucky State University, a historically black teaching institution. These programs will provide the minority students from the teaching university with great opportunities to participate in advanced research activities and build a pipeline for these minority students to enter graduate programs in STEM. (3) The research outcome derived from this project will be integrated into the investigators’ teaching courses such as EE-599 Introduction to Robotics and EE-518 Electric Motor Drives at the University of Kentucky. This will be highly bene?cial for the professional training of future scientists and engineers. (4) The undergraduate and graduate students on this project will receive dedicated training and technical supervision through the project execution, which will directly contribute to the local and regional workforce development once these STEM students graduate and join the industries. (5) The team will disseminate the ?ndings to the national robotics community, through publications and presentations in ?agship international conferences and journals, as well as organize webinars and symposiums related to this topic. 1