Abstract: To address the low visualization in deep cavity structure assembly, which affects assembly quality, a multi-sensor fusion approach for real-time sensing and precise control is proposed. First, the assembly characteristics of deep cavity structures are analyzed, with geometric features and assembly loads identified as key factors affecting assembly quality. The functional modules and logical relationships of the assembly system are introduced, combining digital twin theory. Next, a sensing study of the assembly process is conducted, utilizing laser and force sensors to effectively capture the geometric and load features of the deep cavity structure. Finally, using assembly feature information as input, virtual assembly and error compensation research is carried out. Through assembly path planning and robot end-effector compliance control, precise regulation of the assembly process is achieved. A sensing and control system is developed based on a digital twin framework. Key technology validation on a specific aircraft’s deep cavity structure demonstrates that this method can effectively sense and control multi-stage feature information, providing an effective theoretical framework and technical solution for high-performance assembly.

Key words: deep cavity structure, multi-sensor fusion, assembly process sensing, virtual assembly, assembly control