Universal Learning Approach of an Intelligent Algorithm for Non-GNSS Assisted Beamsteering in V2I Systems

In intelligent transportation systems, an important task is to provide a highly efficient communication channel between vehicles and other infrastructure objects that meets energy efficiency requirements involves low time delays. The paper presents method for generating synthetic data of the “vehicle-to-infrastructure” system, capable simulating many scenarios traffic situations increase generalizing ability beamsteering algorithm. algorithm based on gradient boosting designed connect track with minimal delays without relying GNSS coordinates. predictors applied machine learning were: relief, vehicle type, direction movement speed, timestamps, received signal power level. generated dataset included model Lighthill–Whitham–Richards macroscopic SUMO software package simulations. Simulation results showed 94% accuracy in correctly identified positions main lobe according behavior.