Wearable Collaborative Continuum Robot for Manufacturing Aiding

Wearable collaborative robot (WCR) can serve as a "third hand" for humans, offering potential advantages in assisting to complete manufacturing tasks that cannot be done by only one person. In consideration of stringent requirements for manufacturing scenarios, developing a WCR with outstanding comprehensive performances in terms of safety, agility, precise operation, and strong load capacity remains a formidable challenge. Thus, this paper proposes a wearable collaborative continuum robot (WCCR) for manufacturing aiding through the integrated camera and gripper. The WCCR is designed based on hybrid driving with ropes and pneumatic actuation, endowed with bending, extensile, and variable stiffness capabilities. Concurrently, the control strategy based on a deep reinforcement learning model is proposed to improve the robot's operating abilities without additional sensors. The experimental results indicate that the proposed WCCR has natural safety, flexibility, and excellent assisting working abilities with a high load capacity of 19.8 N, a high stiffness capacity of 140.4 N/m, and an elongation rate of 53$\%$. Under changeable load conditions, the proposed control approach yields an average motion precision of 9 mm.