Secure VANET Communication for Malicious Node Detection using a Classifier Model
Vehicular ad-hoc networks (VANETs) are vulnerable to various attacks because of its dynamic in nature. The significance in the VANET is to estimate the trustworthiness of a vehicle for improving communication security. The several works have been done in the secure communication between the vehicular nodes. But accurate attack detection is still challenging to improve communication security in VANET. In order to improve secure communication in VANET with minimal time consumption, a Multi-Objective Reweighted Adaptive Boosting Classifier based Attack Detection technique is introduced. Reweighted Adaptive Boosting Classifier is an ensemble classifier which creates several weak learners. For classifying the vehicle nodes as normal or attack, the artificial neural network (ANN) is employed as a weak learner. Weak learner includes three layers namely input layer, hidden layer and output layer. Input layer obtains number of vehicle nodes. In hidden layer, multiple objective functions like trust, energy and cooperativeness of the vehicle nodes are computed. The Gaussian activation function is used at the output layer to classify the node as normal or attack based on the multiple objective functions of the vehicle nodes. At last, the approach combines the weak classifier output and offers strong classification results with lesser error rate. Therefore, the attack nodes are detected with minimal energy consumption. Simulation is conducted with metrics namely detection rate (DR), false positive rate (FPR) and detection time (DT) with number of vehicular nodes.