TunnelLuxNet: A unified and fast conditional CNN-Transformer framework for 3D illuminance field prediction in road tunnels
Authors:
- Tao WangTWTao Wang
- Yingying ZhangYZYingying Zhang
- Qiang ZengQZQiang Zeng
- Yushuai ZhaoYZYushuai Zhao
- Junhao XuJXJunhao Xu
Preprint
- 17 October 2026
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
Comprehensive tunnel lighting assessment requires three-dimensional illuminance distributions across the roadway and surrounding surfaces. Physically based simulation resolves these distributions but becomes costly during repeated evaluations, while data-driven models commonly focus on individual planes. Here we present TunnelLuxNet, a conditional CNN–Transformer framework for rapid prediction of tunnel interior illuminance fields. A geometry-aware encoding maps connected surfaces to a common layout represented by multichannel spatial fields and a condition vector. Within this representation, convolutional processing captures local features, axial attention establishes longitudinal and cross-surface interactions, and conditional modulation incorporates geometric, material and luminaire conditions. Trained on 3,000 Radiance-generated configurations, TunnelLuxNet achieves mean MAE and RMSE of 0.097 and 0.404 lx for local fields and 1.290 and 2.286 lx under the densest tested installation, with close agreement against published measurements. Predicting approximately 1.8 million surface illuminance values for a long tunnel takes 22 s, compared with an estimated 6 h using Radiance. Extrapolation tests reveal greater sensitivity to cross-sectional geometry and luminaire aiming than to reflectance or alignment. Separately trained models for batwing and narrow-beam photometries demonstrate the workflow’s applicability to different luminaires. Ablation results highlight cross-surface information exchange as a key contribution to accuracy, suggesting its value for learning coupled physical fields in three-dimensional environments. TunnelLuxNet provides a reusable computational basis for comprehensive lighting assessment and extensive evaluation of design alternatives.Keywords
This publication has 20 references indexed in Scilit:
- Real-time dimming control of integrated daylight and electric lighting systems using a multimodal generative surrogate modelBuilding and Environment, 2026
- Daylight factor prediction using machine learning: A two-way study using numerical encoding and regression models, versus image encoding and pix2pixBuilding and Environment, 2025
- Fusing Transformer and diffusion for high-resolution prediction of daylight illuminance and glare based on sparse ceiling-mounted inputBuilding and Environment, 2024
- Coupling convolutional neural networks with gated recurrent units to model illuminance distribution from light pipe systemsBuilding and Environment, 2023
- Optimal tunnel lighting design in aspect of lighting quality and energy performanceTunnelling and Underground Space Technology, 2022
- Holistic digital-twin-based framework to improve tunnel lighting environment: From methodology to applicationBuilding and Environment, 2022
- An empirical investigation of disability glare and visibility level during driving inside very long road tunnels: A case studyTunnelling and Underground Space Technology, 2022
- Diffuse reflection-based lighting calculation model and particle swarm optimization algorithm for road tunnelsTunnelling and Underground Space Technology, 2022
- Determining multidimensional diffuse reflection effects in city tunnel lighting environmentBuilding and Environment, 2022
- Development and validation of a Radiance model for a translucent panelEnergy and Buildings, 2006