Active self-protection control strategy of damaged robot joints

Key Points

• The aim is to develop a control strategy that allows robots to autonomously recover from joint damage while preventing further deterioration.
• Introduced a health factor to quantify the torque capacity of damaged joints.
• Implemented a three-layer control architecture.
• First layer allocates tasks to healthy joints using weighted pseudo-inverse and null space reconstruction.
• Second layer uses linear time-varying model predictive control to impose dynamic constraints.
• Third layer involves a closed-loop feedback mechanism based on dynamic feedforward.
• With damaged joints having only 30% torque capacity, the robot maintained its planned trajectory.
• Torque on damaged joints was limited to safe levels, preventing further deterioration.
• The strategy improved operational reliability under conditions of joint impairment.