A Human-Computer Collaborative Behavior Measurement Model for Assembly in Constrained Visibility Environments

Collision interference detection is an important concern in the manual installation of cable harnesses. Due to the complex and variable layout of cable harness installations, hand assembly movements are prone to collisions, contact, and other forms of interaction with surrounding equipment. Furthermore, cable harnesses often need to be routed along specific paths in visually constrained environments. Currently, there is a lack of modeling in the extraction of hand motion parameters and the data analysis of hand action intent. To address this challenge, we propose a novel human-machine collaboration behavior measurement model. This model not only provides a rapid solution for extracting hand motion parameters but also delivers efficient and natural visual feedback for the behavioral intent reflected by hand motion features. First, we introduce a hand motion parameter extraction mechanism based on a hand kinematics model. Second, we develop a virtual-to-real spatial registration model specifically designed for visually constrained conditions, enabling accurate recognition and 3D calibration of hand action intent. A user study experiment demonstrates that the proposed model outperforms traditional hand behavior measurement models in terms of manual task efficiency, hand motion recognition accuracy, and the naturalness of hand interactions. This improvement is particularly evident in visually constrained environments, effectively addressing challenges in obstacle avoidance and intent inference during spatially constrained assembly tasks.OPEN ACCESS Received: 10/09/2024 Accepted: 08/11/2024 Published: 20/04/2025