Abstract: Conventional tunnel lighting systems face challenges such as high energy consumption and limitations in dynamically coordinating real-time traffic flow with luminance parameters inside/outside tunnels for energy-saving control. To address these challenges, the authors propose a real-time tunnel lighting regulation system based on multi-source data fusion. First, a heterogeneous data fusion architecture combining millimeter-wave radar and luminance detectors is designed, achieving millisecond-level synchronous perception of traffic volume, vehicle speed, and intra/extra-tunnel luminance. Subsequently, an edge-cloud collaborative dimming model is constructed, where edge computing nodes execute lightweight algorithms to generate real-time dimming commands, while cloud-based optimization of threshold parameters is performed through long-term data training. Finally, a regional "on-demand real-time dimming" strategy is proposed to adaptively match baseline illumination with exterior luminance. A management platform integrating multi-source data analysis, three-dimensional visualization, and lighting status monitoring ensures real-time tunnel safety. Experimental deployment in the left tunnel of a Yunnan province case study demonstrates that the system achieves energy-saving ratios of 63.90%－67.43% compared to time-sequence circuit control under high traffic volumes (>1 500 vehicles per day). During low-traffic periods (≤1 500 vehicles per day), energy-saving efficiency exhibits a negative correlation with traffic volume, reaching up to 88.43%, while fully complying with tunnel operational safety standards.

Key words: highway tunnel, real-time dimming system, lighting control strategy, vehicle-adaptive demand control, energy-efficiency management