Autonomous Fault-Tolerant Operation of Redundant Robotic Arms

点击翻译按钮获取中文摘要

Environments with extreme temperatures, humidity, radiation, and other hazardous conditions often escalate the potential risk of hardware failures of robot actuators. A common root cause contributing to the frequent robot joint failures is the faults with the joint motors and the associated servo drives. Unfortunately, in unstructured, remote and dangerous environments, such as when robots are deployed for space exploration, nuclear waste remediation, and disaster rescue, not only are failures more likely to occur, but it is also impossible to perform routine maintenance for these robots after a failure occurs. Failures occurring in other safety-critical applications, such as robotic surgery, rehabilitation, and human-robot interaction, could also lead to very serious consequences or accidents. To improve robustness, this project will develop autonomous fault-tolerant and fail-active strategies for redundant robot arms based on predicted and identified joint faults, which will dramatically improve the reliability and robustness of redundant robot arms.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 occur. The kinematic design of optimally fault-tolerant robots and fault-tolerant motion planning methods in anticipation of all potential failures can guarantee task completion and optimal post-failure performance. In particular, a novel efficient method is planned to compute the six-dimensional fault-tolerant workspace that includes both the volume and shape information. Then, the optimally fault-tolerant robots will be deigned based on the volume and shape of the fault-tolerant workspace. Additionally, existing fault-tolerant motion-planning algorithms assume that all the joints have equal probability to fail, and they can only provide local optimal solutions. In this research, the post-failure performance measures are developed based on predicted joint failure probabilities, and a real-time motion planning algorithm will be introduced to compute a globally optimal trajectory. Finally, the prognostic and diagnostic algorithms differ from all the conventional approaches in that it does not require any additional sensors or hardware, but still maintains high accuracy and fast prognostic and diagnostic speed.This project is supported by the cross-directorate Foundational Research in Robotics program, jointly managed and funded by the Directorates for Engineering (ENG) and Computer and Information Science and Engineering (CISE). This project is also jointly funded by the Established Program to Stimulate Competitive Research (EPSCoR).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.