Positioning System Design for Power Construction Sites Using Multi-Sensor Fusion

Aiming at the urgent demand for high-precision positioning in the full indoor-outdoor domain at power construction sites, this paper designs a positioning system based on the “RTK+UWB+IMU” heterogeneous fusion architecture. The system achieves outdoor centimeter-level positioning through RTK and adopts a CNN-BILSTM deep learning model to optimize the positioning accuracy of UWB in complex indoor environments. It innovatively constructs a fuzzy rule base to realize seamless switching between indoor and outdoor positioning based on indicators such as GNSS satellite signal quality and UWB base station signal strength, and combines IMU to enhance the positioning robustness in transition areas. The supporting digital management platform integrates functions of real-time positioning tracking, intelligent electronic fence alarm, and historical trajectory retrieval. Experimental results show that the system’s static positioning error is less than 10 cm both indoors and outdoors, and the trajectory in transition areas is smooth without jumps, providing intelligent technical support for safety management and control of power construction.