Reliable data exchange in vehicular delay tolerant networks using AI-based hybrid trust management

Vehicular delay-tolerant networks (VDTNs) play prominent part in smart transportation and smart city paradigm to diminish the drivers' risk to meet accidents and offer many infotainment and entertainment services. All the operations are carried out by exchanging data among nodes. However, due to open communication nature of VDTN, the data broadcasted by mobile nodes are quiet susceptible to security attacks. Recently, a number of protocols are proposed for open communication. However, securing such communications and establishing trust value among vehicles are major addressable issues. This is because the existing fraudulent peers can lead to catastrophic circumstances on roads. Along with securing the data carriers, the authenticity and reliability of data is also equally important. In order to identify reliable nodes and maintaining the authenticity of data being communicated, a new artificial intelligence (AI)-based hybrid trust management framework is proposed in this article. The study particularly exploits the proximity-based k-nearest neighbour (kNN) classification model for selecting controlling nodes. Further, decision stump is employed to accomplish trust estimation rules, and backpropagation is used to self-train the mobile nodes, whenever anticipated trust value is not attained. The proposed scheme maintains the integrity of vehicular nodes and efficiently handles the specified. The trust framework uses a multifaceted direct and recommended trust estimation approach to calculate the global trust values. For comparison, the performance of the proposed method was compared with other trust management approaches.

Automated summary

This article proposes an artificial intelligence-based hybrid trust management framework to protect the security of vehicular delay-tolerant networks (VDTNs) and ensure the authenticity and reliability of data. By using a proximity-based k-nearest neighbour classification model to select controlling nodes and employing decision stumps for trust estimation rules, the proposed scheme can effectively handle the issues of vehicular nodes.