Consensus-Decoupling Based Task Planning: A Positive Synergy Approach for a Human-In-the-loop Collaboration System

Human-In-the-loop collaborative systems constitute a critical focus area in Industry 5.0, where the rationality of task planning and collaborative allocation in front-end processes directly impacts the quality-efficiency level of human-machine coupling collaboration. Current research has paid limited attention to the global optimization decomposition of complex human-machine collaborative tasks, thereby constraining the optimization space for subsequent collaborative allocation. This study proposes a human-machine collaborative coupling task partitioning method based on multi-view consensus decoupling computation. Integrating perspectives of human-machine responsibility decomposition optimization and robustness under uncertain task environments, we establish a decoupling criterion comprising three dimensions: interaction dependency, quality-efficiency correlation, and adaptability of granularity. A robust matrix factorization with spectral embedding method considering multi-view perspectives is employed to optimize the consensus clustering partitioning scheme. The proposed methodology is validated through human-machine collaborative task scenarios in wind turbine cluster maintenance operations under mountainous environments, demonstrating its feasibility and effectiveness.