ROBOT-BASED DYNAMIC RESTORATION DESIGN FOR INDUSTRIAL CULTURAL HERITAGE PROTECTION AND MAINTENANCE USING SENSING TECHNOLOGY

Yanpeng Li and Guofeng Zhang

Keywords

Industrial cultural heritage, robotic system design, multimodalsensor fusion∗ School of Architecture and Art Design, University of Scienceand Technology Liaoning, Anshan 114001, China; e-mail:[email protected]; [email protected] author: Yanpeng LiRecommended by Zhuming Bi

Abstract

To address heterogeneous damage in industrial cultural heritage structures under complex environmental conditions, this paper proposes a dynamic restoration system based on robotic and multimodal sensing technologies. The system integrates a 6-degree- of-freedom robot arm with thermal imaging, LiDAR, and surface vibration sensors, and establishes a closed-loop control architecture for “identification-planning-execution-feedback”. A YOLOv8+ UNet composite model is proposed to detect the target, and a modified RRT algorithm is used to optimise the repair path. The control system adopts a hybrid vision-servo and force-position control mechanism, validated through both simulations and physical platform testing. Experimental results show that the recognition precision for linear cracks reaches 94.3%, and the average path execution error is controlled within ±0.74 mm. Under the most complex multi-segment combined path, the maximum execution error reaches ±1.26 mm, while the medium-frequency control layer maintains a real-time adjustment amplitude of ±0.82 mm. Compared with manual operation, single task repair time is reduced by 47.3%. The experimental results demonstrate the adaptability and feasibility of robotic sensor fusion for the restoration of complex industrial heritage structures.

Important Links:

Go Back