6G AND V2X: REVOLUTIONING CONNECTED VEHICLE NETWORKS

Vehicle-to-Vehicle (V2V) communication isessential for improving road safety and optimizing trafficflow. Machine learning techniques, such as Support VectorMachines (SVM), offer an efficient method for estimatingcommunication reliability in real-time without requiring acomplete network model. In this study, Ant ColonyOptimization (ACO) was applied to enhance the trainingprocess of an SVM model for classifying V2V messagereliability. The model underwent ten-fold crossvalidationbefore being implemented using Verilog HardwareDescription Language (VHDL) for FPGA execution,enabling a cost-effective and efficient hardware solution. Byutilizing a linear SVM, the system achieved low powerconsumption, using only 1.4% of the Xilinx Artix7 FPGA’sresources. This efficiency allows multiple instances of thereliability estimation model to operate simultaneously,facilitating the monitoring of multiple vehicle links.