30 April 2021 in IEEE Transactions on Vehicular Technology
The vehicular networks constructed by interconnected vehicles and transportation infrastructure are vulnerable to cyber-intrusions due to the expanded use of software and the introduction of wireless interfaces. Intrusion detection systems (IDSs) can be customized efficiently in response to this increased attack surface. There has been significant progress in detecting malicious attack traffic using machine learning approaches. However, existing IDSs require network devices with powerful computing capabilities to continuously train and update complex network models, which reduces the efficiency and defense capability of intrusion detection systems due to limited resources and untimely model updates. This work proposes a cooperative intrusion detection mechanism that offloads the training model to distributed edge devices (e.g., connected vehicles and roadside units (RSUs). Distributed federated-based approach reduces resource utilization of the central server while assuring security and privacy. To ensure the security of the aggregation model, blockchain is used for the storage and sharing of the training models. This work analyzes common attacks and shows that the proposed scheme achieves cooperative privacy-preservation for vehicles while reducing communication overhead and computation cost.
27 January 2020 in IEEE Transactions on Vehicular Technology
27 August 2021 in IEEE Transactions on Neural Networks and Learning Systems
Android is undergoing unprecedented malicious threats daily, but the existing methods for malware detection often fail to cope with evolving camouflage in malware. To address this issue, we present Hawk, a new malware detection framework for evolutionary Android applications. We model Android entities and behavioral relationships as a heterogeneous information network (HIN), exploiting its rich semantic meta-structures for specifying implicit higher order relationships. An incremental learning model is created to handle the applications that manifest dynamically, without the need for reconstructing the whole HIN and the subsequent embedding model. The model can pinpoint rapidly the proximity between a new application and existing in-sample applications and aggregate their numerical embeddings under various semantics. Our experiments examine more than 80,860 malicious and 100,375 benign applications developed over a period of seven years, showing that Hawk achieves the highest detection accuracy against baselines and takes only 3.5 ms on average to detect an out-of-sample application, with the accelerated training time of 50x faster than the existing approach.
12 November 2019 in IEEE Transactions on Industrial Informatics
Empowered with urban informatics, transportation industry has witnessed a paradigm shift. These developments lead to the need of content processing and sharing between vehicles under strict delay constraints. Mobile edge services can help meet these demands through computation offloading and edge caching empowered transmission, while cache-enabled smart vehicles may also work as carriers for content dispatch. However, diverse capacities of edge servers and smart vehicles as well as unpredictable vehicle routes make efficient and time content distribution a challenge. To cope with this challenge, we exploit the relations between vehicles and road side units in content dispatch, and develop a social-aware mobile edge computing and caching mechanism. By leveraging deep reinforcement learning approach, we propose optimal content processing and caching schemes that maximize the dispatch utility in an urban environment with diverse vehicular social characteristics. Numerical results based on real urban traffic datasets demonstrate the efficiency of our proposed schemes.
11 July 2022 in IEEE Transactions on Vehicular Technology
To deal with the security and privacy issues in vehicular ad hoc network (VANET), digital signature based on public key infrastructure (PKI) is recognized as being the most mature and popular approach. However, existing secure and privacy-preserving schemes for vehicular communications face some challenges, e.g., fast verification of signatures and certificates, reducing the storage/transmission overheads, avoiding the dependence on ideal tamper-proof devices. In this paper, we propose a hybrid and hierarchical aggregation-verification scheme for efficient verification of certificates and signatures on safety messages in VANET. In our scheme, short-term anonymous certificates and aggregate signatures are employed to guarantee the security and privacy of the VANET system. A novel hybrid aggregation-verification technique is proposed to firstly aggregate the signatures in different certificates and signatures on safety messages into fixed length aggregate signature and then verify the aggregate signature with very short delay. To the best of our knowledge, this paper achieves signature aggregation and verification of such kind for the first time. Further, our scheme also enables hierarchical aggregation-verification of certificates and signatures in VANET.
11 August 2023 in IEEE Transactions on Intelligent Vehicles
The scenario-based tests have been an important means of improving the safety and reliability of autonomous vehicles (AVs). Moreover, ensuring the fidelity of generated scenarios is a safeguard for scenario-based tests. This paper presents a new approach to generate twin scenarios for the virtual testing of AVs using meta scenarios. By extracting key parameter clusters from real-world scenarios and redefining and reassembling them, a large number of twin scenarios are generated. This approach addresses the need for diverse scenarios in AV virtual testing and enhances efficiency and hazard detection in the Safety of the Intended Functionality (SOTIF) assessment. The validity of twin scenarios, their temporal logic, and social attributes are verified using the UPPAAL tool, ensuring reliability and fidelity in virtual testing with AV digital twins. This innovative method offers a valuable solution to address challenges in AV virtual testing and improve the overall assessment process.