Solar-powered VANET architecture for sustainable and smart traffic management in arid regions

Abstract Energy-autonomous roadside communication is essential for sustainable intelligent transportation systems (ITS), particularly in arid regions that combine abundant solar resources with highly dynamic traffic demand. Existing solar-powered vehicular ad hoc network (VANET) approaches typically treat roadside unit (RSU) placement, activation scheduling, and traffic prediction as separate problems, limiting their ability to coordinate long-term infrastructure siting with short-term communication demand and energy availability. This paper presents an integrated solar-powered VANET framework that coordinates these spatial and temporal decisions within a unified architecture. A Discrete Grey Wolf Optimization (D-GWO) method performs spacing-aware RSU siting by maximizing traffic coverage, network centrality, and solar harvesting potential while penalizing redundancy. A Traffic-Aware Duty Cycling (TADC) controller regulates RSU operation through traffic hysteresis, battery state-of-charge guards, and minimum dwell times. A Joint Activation Score (JAS) couples long-term site utility with runtime activation, and a Random Forest-based traffic anticipation module enables proactive scheduling; the Activation Timeliness Gain (ATG) metric quantifies the resulting improvement in activation timing. Trace-driven simulations on the Ha’il City road network, using measured solar irradiance and real traffic data, demonstrate that the proposed framework achieves packet delivery ratios (PDR) of 94.5–97.9%, reduces end-to-end delay (E2ED) by 7–35 ms, and lowers per-RSU energy consumption by 20–33% relative to representative baselines, while maintaining fully off-grid operation under the evaluated conditions. Predictive assistance further advances RSU activation by up to 2.1 min during traffic surges without additional energy cost. These results show that coordinated spatio-temporal design can simultaneously enhance communication reliability, responsiveness, and energy efficiency, supporting the feasibility of energy-autonomous VANET operation under the evaluated Ha’il topology, traffic densities, irradiance trace, and log-distance radio model.