Dynamic Intrusion Detection Framework for UAVCAN Protocol Using AI

Industry 4.0 is going through a transitional period via the radically automotive transformations. In particular, unmanned aerial vehicles have significantly contributed to the development of intelligent and connected transportation systems. Thus, the continuous development using diverse technologies to achieve a variety of high-performance services raised the security concerns regarding communicating entities. Thus, being managed by networked controllers, UAVs uses controller area networks (CAN) protocol to broadcast information in a bus. However, this protocol is used as a de facto standard which does not have sufficient security features that raise the security risks. This issue caught the attention of the automotive industry researchers and several studies have attempted to improve the security of the CAN protocol attack detection. However, the proposed studies established general perspective solution and did not pay attention to UAVCAN attack detection. To alleviate these concerns, this paper proposed a dynamic intrusion detection frameworks (DIDF) for UAVCAN. The proposed UAVCAN DIDF scheme adopts an artificial intelligence (AI) based model to achieve high detection performance. We performed experiments using public UAVCAN dataset to evaluate our detection system. The experimental results demonstrate that UAVCAN DIDF has significantly reached a high detection rate with a high true positive and a low false negative rate. The simulation results are encouraging and demonstrate the effectiveness of UAVCAN DIDF.