The development of autonomous vehicles in compliance with safety requirements requires them being equipped with various types of sensors. This integration presents significant potential for enhancing communication performance within Vehicular Ad-hoc Networks (VANETs). A critical communication requirement in VANETs is achieving low delay to enable realtime communication and reducing routing overhead to increase the generation rate of safety-related packets, thereby enhancing vehicular safety. This paper aims to address this issue by leveraging sensor data through proposing an optimized packet routing method within the network. To establish a common structure among different vehicles, roadways are divided into smaller segments with adaptive lengths, and sensor information from each occupied segment is disseminated within a shared framework inspired by 5G sidelink using a Time Division Integrated Sensing and Communication (TD-ISAC) frame structure. The dimensions of this frame are optimized using a combined Lagrangian Relaxation and Branch and Bound approach. Additionally, vehicle positions are estimated in real-time based on sensor data through the integration of a Kalman filter and a Transformer model, resulting in the formation of radial awareness in the network topology. Finally, the relay station selection process is stabilized by incorporating reserve relay stations. The proposed model is evaluated using SUMO and compared to the Zone Routing Protocol (ZRP) and Ad Hoc On-Demand Distance Vector (AODV) algorithms, demonstrating up to a 60% improvement in delay and a 5% reduction in routing overhead under multi-hop scenarios.