No abstract

Multi-Target tracking is a central aspect of modeling the environment of autonomous vehicles. A mono camera is a necessary component in the autonomous driving system. One of the biggest advantages of the mono camera is it can give out the type of vehicle and cameras are the only sensors able to interpret 2D information such as road signs or lane markings. Besides this, it has the advantage of estimating the lateral velocity of the moving object. The mono camera is now being used by companies all over the world to build autonomous vehicles. In the expressway scenario, the forward-looking camera can generate a raw picture to extract information from and finally achieve tracking multiple vehicles at the same time. A multi-object tracking system, which is composed of a convolution neural network module, depth estimation module, kinematic state estimation module, data association module, and track management module, is needed. This paper applies the YOLO detection algorithm combined with the depth estimation algorithm, Extend Kalman Filter, and Nearest Neighbor algorithm with a gating trick to build the tracking system. Finally, the tracking system is tested on the vehicle equipped with a forward mono camera, and the results show that the lateral and longitudinal position and velocity can satisfy the need for Adaptive Cruise Control (ACC), Navigation On Pilot (NOP), Auto Emergency Braking (AEB), and other applications.

In the context of the evolution of in-vehicle electronic and electrical architecture as well as the rapid development of quantum computers, post-quantum algorithms, such as NTRUEncrypt, are of great significance for in-vehicle secure communications. In this paper, we propose and evaluate, for the first time, a NTRUEncrypt enhanced session key negotiation for the in-vehicle Ethernet context. Specifically, the time consumption and memory occupation of the NTRUEncrypt Elliptic Curve Diffie–Hellman (ECDH), and Rivest–Shamir–Adleman (RSA) algorithms, which are used for session key negotiation, are measured and compared. The result shows that, besides the NTRUEncrypt’s particular attribute of resisting quantum computer attacks, the execution speed of session key negotiation using NTRUEncrypt is 66.06 times faster than ECDH, and 1530.98 times faster than RSA at the 128-bit security level. The memory occupation of the algorithms is at the same order of magnitude. As the transport layer security (TLS) protocol can fulfill most performance requirements of the automotive industry, post-quantum enhanced session key negotiation will probably be widely used for in-vehicle Ethernet communication.

The automotive Ethernet is gradually replacing the traditional controller area network (CAN) as the backbone network of the vehicle. As an essential protocol to solve service-based communication, Scalable service-Oriented MiddlewarE over IP (SOME/IP) is expected to be applied to an in-vehicle network (IVN). The increasing number of external attack interfaces and the protocol's vulnerability makes SOME/IP in-vehicle networks vulnerable to intrusion. This paper proposes a multi-layer intrusion detection system (IDS) architecture, including rule-based and artificial intelligence (AI)-based modules. The rule-based module is used to detect the SOME/IP header, SOME/IP-SD message, message interval, and communication process. The AI-based module acts on the payload. We propose a SOME/IP dataset establishment method to evaluate the performance of the proposed multi-layer IDS. Experiments are carried out on a Jetson Xavier NX, showing that the accuracy of AI-based detection reached 99.7761% and that of rule-based detection was 100%. The average detection time per packet is 0.3958 ms with graphics processing unit (GPU) acceleration and 0.6669 ms with only a central processing unit (CPU). After vehicle-level real-time analyses, the proposed IDS can be deployed for distributed or select critical advanced driving assistance system (ADAS) traffic for detection in a centralized layout.