Deep Spiking Neural based Dolphin Echolocative Node Optimized Bayesian Regression For Resource-Efficient and Load-Balanced Data Transmission in VANET

Vehicular Ad-hoc Networks (VANETs) have potential to completely transform transportation infrastructure with its increased safety, effectiveness, and convenience. VANETs are characterized by rapidly changing topologies as nodes i.e., vehicles move in and out of communication range. In order to increase the performance in communication, a vehicle must provide its data to other vehicles, which calls for sufficient resources. The load balancing helps to avoid congestion and enhance network performance in general, through the equitable distribution of traffic among available nodes. For that reason, this research is motivated by the Bayesian Dolphin Echolocative Node Optimized Deep Spiking Neural Network (BDENODSNN) method to provide efficient resource allocation with load balancing for ensuring robust data communication in VANET. Different number of vehicle nodes is taken for input layer. Following that, first hidden layer utilizes Dolphin Echolocative Node Optimization process for discovering resource-efficient vehicle nodes in VANETs. Here, the vehicle nodes with higher residual energy and fewer loads are selected as optimal vehicle nodes. After this, the optimal nodes are transmitted to the second hidden layer whereas the regression process is carried out by performing three processes such as estimation of load capacity, identification of less loaded and heavy loaded vehicles nodes and Nearest Neighbour Search utilizing Bayesian Linear Regression. Here, the nearby vehicle node with low load at the shortest distance is selected that helps in establishing data transmission with higher throughput and less delay.