XGBoost-deep neural network fusion model for network selection-based vertical handover decision making and optimized routing for heterogeneous VANET

With the aid of efficient network selection and routing, this research suggests a unique vertical handoff/handover (VHO) mechanism for Heterogeneous wireless networks (HWN). The three stages of this paradigm are handover execution, handover decision-making, and handover triggering. Initially, the handover triggering phase, which calculates the vehicle's Travelling Distance (TD) using Enhanced Extended Kalman Filter (E 2 KF) and RSS of the currently serviced network in the coverage area using Long Short-Term Memory (LSTM) is initiated. Secondly, the handover decision-making system is performed optimally by selecting the appropriate network for handover using the proposed XGBoost-Deep Neural Network Fusion Model (XGDN-FM). Finally, after the handover decision phase, the handover execution is carried out by selecting optimal Vehicle 2 Vehicle (V2V) routing using the proposed Trend factor Smoothing Brown Bear Optimization Algorithm (TS-BBOA). The outcomes showed that the model attains the lowest PLR of 0.18%, highest throughput of 4.89Mbps, low energy consumption of 2.4 J, and highest handover success probability with less handover failure and unnecessary handovers.