A Novel Approach for Dynamic Stable Clustering in VANET Using Deep Learning (LSTM) Model

Abstract

Clustering in VANETs, which dynamically evolve into wireless networks, is difficult due to the networks' frequent disconnection and fast changing topology. The stability of the cluster head (CH) has a huge impact on the network's robustness and scalability. The overhead is decreased. The stable CH assures that intra- and inter-cluster communication is minimal. Because of these difficulties, the authors seek a CH selection technique based on a weighted combination of four variables: community neighborhood, quirkiness, befit factor, and trust. The stability of CH is influenced by the vehicle's speed, distance, velocity, and change in acceleration. These are considered for in the befit factor. Also, when changing the model, the precise location of the vehicle is critical. Thus, the predicted location is used to evaluate CH stability with the help of the Kalman filter. The results showed that the befit factor performed better than the latest developments. Because of the high speed of the vehicle, dynamic changes and frequent communication link breaks are unavoidable. In order to fully perceive issue, a graphing approach employed to assess the eccentricity then the communal neighborhood. Using Eigen gap heuristic, the link dependability is determined. Trust is the final important parameter that has not yet been taken into account in the weighted method. The trust levels are specifically being evaluated for the primary users using an adaptive spectrum sensing. Long short-term memory (LSTM), a deep recurrent learning network, used to train the likelihood of detection under diverse signal and noise situations. By using LSTM model, significantly decreased the false rate. The cluster head stability has improved for high traffic density, significantly improved according to the comparative analysis with the weighted and individual metrics. The efficiency of the network has also greatly increased in terms of throughput, packet delay, packet delay ratio, and energy consumption.