Abstract—The rapid expansion of vehicular communication systems emphasizes the integration of LTE-V networks, crucial for applications like road safety, traffic management, and infotainment. High-speed scenarios demand efficient downlink scheduling due to constantly changing channel conditions influenced by factors like throughput and Bit Error Rate (BER). Mobility-induced channel variations lead to signal quality fluctuations, interference, and congestion. LTE-V networks require robust Quality of Service (QoS) for safety applications, necessitating algorithms that detect and mitigate interference by dynamically adjusting scheduling. Existing algorithms struggle with Doppler shift effects, interference, and predicting network patterns, prompting the exploration of an Intelligent Downlink Scheduling (IDS) scheme based on Support Vector Machines (SVM) for highspeed LTE-V networks. This work focuses on the optimization of the resource allocation, improving spectral efficiency, and predicting network congestion. Leveraging machine learning and optimization, it addresses challenges posed by varying vehicle densities, mobility patterns, and QoS needs. Extensive simulations show the IDS’s superiority, significantly enhancing throughput and reducing BER. The improved throughput signifies reduced data loss in scheduling queues, while lower BER indicates enhanced received data post-scheduling. The IDS facilitates real-time decision-making and data-driven insights, ideal for managing and optimizing downlink scheduling in dynamic Long-Term Evolution-Vehicle (LTE-V) networks. Simulation results demonstrate a substantial 13 dB improvement over the best CQI scheduler at a 10-4 BER and a 24 Mbps increase at a 20 dB SNR for a vehicle density of 40, showcasing the IDS's performance enhancements.
 

Keywords—Long-Term Evolution-Vehicle (LTE-V), intelligent downlink scheduling, vehicular communication, machine learning, spectral efficiency, quality of service