Optimized energy efficient cluster based routing protocol for real time communication in 5G enabled VANETs using reinforcement learning

The fusion of fifth generation (5G) communication technologies with Vehicular Ad Hoc Networks (VANETs) provides the possibility to achieve low latency and reliable communication for intelligent transportation systems. But with high mobility of the vehicles, the frequent topology change, variable link quality, growing network density, and energy limitation, the traditional routing protocols are facing major challenges. These limitations can result in excessive control overhead, route instability, increased communication delay, and inefficient energy utilization. To overcome the above issues, this study presents an Optimized Energy-Efficient Cluster Based Routing Protocol (OEECRP) for real-time communication in 5G based VANETs by adopting Reinforcement Learning (RL) technique. The suggested framework combines the weighted dynamic clustering, energy and mobility based cluster-head selection, energy-adaptive routing with Q-learning and 5G quality of service. The cluster heads are selected by a weighted process that calculates the residual energy (40%), mobility stability (30%), and 5G link quality (30%). The Q-learning algorithm assesses network states such as vehicle density, battery level, link stability and communication delay for making intelligent routing decisions. The protocol was tested with simulation experiments in MATLAB R2023a with the 5G Toolbox and Communications Toolbox under urban vehicle network conditions with vehicle densities of 50–200. The performance metrics of OEECRP were compared with that of AODV, DSR, DSDV, GPSR, and LEACH with respect to energy consumption, end-to-end delay, packet delivery ratio (PDR), packet loss ratio and throughput. The results reported show that OEECRP kept the energy consumption in the range of 21.3 ± 0.8 to 45.4 ± 1.6, the end-to-end delay in the range of 38.5 ± 1.2 to 59.3 ± 1.9 ms and PDR in the range of 96.7 ± 0.5 to 89.1 ± 0.9 for the vehicle densities evaluated. Throughput increased from 452.1 ± 12 to 592.6 ± 20 kbps. Based on these results, the use of a hybrid approach between energy-adaptive clustering and reinforcement-learning-based routing for enhancing the VANET communication performance under different network densities seems promising. Practical deployment performance has to be further validated via repeated experiments, real mobility traces and hardware based testbeds.