Manipulator control in collaborative assembly

Remote operation of nuclear processes started with electromechanical telemanipulation over 60 years ago. Today, robotic manipulation is essential to several steps of the nuclear fuel cycle, decontamination and decommissioning operations, and many other related industries. However, not all tasks can be accomplished with a fully remote system. This research focuses on the control of a robotic manipulator in direct contact with human operators to improve the safety and throughput of a precision assembly task. Humans are naturally talented at precise force modulation, using tactile feedback and intuition to assemble complex and fragile parts. Robots, on the other hand, outrank humans in positional precision and strength, especially over long periods. By using a robotic system to offload the weight and reduce the inertia of an object, a human can focus on interaction forces and complex maneuvers to better complete precision assembly and insertion tasks. To investigate the validity of this claim, a custom admittance controller was applied to a passively-balanced collaborative robotic manipulator. Experimental results were collected in a blind precision insertion task with a heavy payload and fragile insertion member. In this pilot study, operator performance was assessed with the manipulator in both active and passive states, a passive mechanical gantry with counterweight for gravity compensation, and no assisting mechanisms. Experimental results indicate improvement in success rates, operation times, and physical effort