Trust-based adversary detection in edge computing assisted vehicular networks

Low-latency requirements of vehicular networks can be met by installing mobile edge hosts that implement computing in the roadside units (RSUs). Adversaries can, however, compromise these RSUs and use them to launch cyber attacks. In this paper, we consider an adversary selectively drops packets or corrupts between RSU passing vehicles. Such strategies would lead a higher number re-transmissions thereby increase latency network, which turn impacts critical delay-sensitive applications like collision avoidance, emergency vehicle warning, etc. We propose trust-based detection systems detect such adversary. Each transmits its uplink downlink trust values about every it has interacted with. These are relayed gateway, where decision will made, via next encountered vehicle. At regular intervals, gateway aggregates obtained from multiple It compares against their respective thresholds classify as benign malicious. also presence malicious vehicles trying deceive system reporting false values. A mechanism is proposed Simulation results generated using MATLAB presented demonstrate performance mechanisms impact adversary's parameters on systems.