A multifunctional fusion sensor is designed

and a multi-modal Transformer with attention-based fusion(MTAF) model for sensor-aware dual-task prediction is proposed.This model aims to achieve rapid and accurate detection of rainfall

ambient illuminance

forward illuminance

temperature and relative humidity

thereby realizing intelligent wiper activation

head-up display(HUD) brightness adjustment

automatic headlights

automatic defogging

automatic window closing and other operations on vehicles

enhancing driving safety and intelligence.Experimental results show that the proposed MTAF model achieves high classification accuracy in both HUD brightness adjustment and rain and fog weather recognition tasks

with the accuracies of 99.26% and 97.41%

respectively.This method provides a feasible path for efficient fusion and modeling of multi-source perception data

and possesses good popularization and engineering application potential.