Integrated optimisation of human-robot collaborative disassembly planning and adaptive evaluation driven by a digital twin

With the continuous development of intelligent manufacturing and human-oriented manufacturing, human-robot collaborative disassembly is becoming a new trend in intelligent remanufacturing. The application of digital twin technology in human-robot collaborative disassembly (HRCD) can significantly increase work efficiency and improve human well-being. Herein, we propose a reference framework for digital twin-driven HRCD planning and adaptive evaluation, which integrates three modules: HRCD digital twin environment construction, HRCD sequence optimisation, and HRCD adaptive evaluation. Subsequently, based on the physiological and psychological fatigue of workers, we establish a planning model with disassembly time and disassembly complexity, and propose an improved heuristic algorithm to determine the task allocation scheme. To enable adaptive evaluation of HRCD strategies, a digital twin-driven kernel point convolution neural network model (DTKPN) and a digital twin-driven Bayesian neural network human posture estimation model (DT-BSHP) are implemented for robot recognition and human pose evaluation. The proposed model can leverage the skills of humans and robots, satisfy ergonomic requirements, improve disassembly efficiency, and reduce disassembly complexity. Finally, the method is applied to a simplified satellite disassembly case. It is shown that the proposed model significantly reduces the disassembly time and complexity and thus the effectiveness and sensitivity of the proposed model are verified.