Protocol misbehavior detection framework using machine learning classification in vehicular Ad Hoc networks

Abstract

A novel approach is proposed to detect protocol misbehavior using state-of-the-art machine learning frameworks and entropy. Nodes in Vehicular Ad Hoc Networks (VANETs) use broadcast protocols to efficiently disseminate safety information, but nodes do not always behave according to the routing protocols. Misbehavior can be caused by a targeted attack, where an attacking vehicle can intentionally send or route malicious packets to harm. Due to the dynamic nature of nodes in VANETs and routing complexity, unintentional misbehavior can also happen due to hardware or software failures in the vehicle. We are not concerned with the targeted attacks, but rather explore how the unintentional misbehavior, which can cause statistical multi-hop routing protocols to operate as basic flooding protocols, can be detected and accurately classified. These methods and detection techniques are based on the IEEE 802.11p MAC layer and weighted p-persistence multi-hop routing protocol. The linear classification was done using the TensorFlow framework and evaluations were performed using the VEINS simulator using the p-persistence broadcast protocol in a US city area.