Federated edge learning (FEEL) technology for vehicular networks is considered as a promising technology to reduce the computation workload while keeping the privacy of users. In the FEEL system, vehicles upload data to the edge servers, which train the vehicles' data to update local models and then return the result to vehicles to avoid sharing the original data. However, the cache queue in the edge is limited and the channel between edge server and each vehicle is time-varying. Thus, it is challenging to select a suitable number of vehicles to ensure that the uploaded data can keep a stable cache queue in edge server while maximizing the learning accuracy. Moreover, selecting vehicles with different resource statuses to update data will affect the total amount of data involved in training, which further affects the model accuracy. In this paper, we propose a vehicle selection scheme, which maximizes the learning accuracy while ensuring the stability of the cache queue, where the statuses of all the vehicles in the coverage of edge server are taken into account. The performance of this scheme is evaluated through simulation experiments, which indicates that our proposed scheme can perform better than the known benchmark scheme.
Connected and autonomous vehicle (CAV) vehicle to infrastructure (V2I) scenarios have more stringent requirements on the communication rate, delay, and reliability of the Internet of vehicles (IoV). New radio vehicle to everything (NR-V2X) adopts link adaptation (LA) to improve the efficiency and reliability of road safety information transmission. In order to solve the problem that the existing LA scheduling algorithms cannot adapt to the Doppler shift and complex fast time-varying channel in V2I scenario, resulting in low reliability of information transmission, this paper proposes a deep Q-learning (DQL)-based massive multiple-input multiple-output (MIMO) LA scheduling algorithm for autonomous driving V2I scenario. The algorithm combines deep neural network (DNN) with Q-learning (QL) algorithm, which is used for joint scheduling of modulation and coding scheme (MCS) and space division multiplexing (SDM). The system simulation results show that the algorithm proposed in this paper can fully adapt to the different channel environment in the V2I scenario, and select the optimal MCS and SDM for the transmission of road safety information, thereby the accuracy of road safety information transmission is improved, collision accidents can be avoided, and bring a good autonomous driving experience.
With the development of artificial intelligence (AI) and 5G technology, the integration of sensing, communication and computing in the Internet of Vehicles (IoV) is becoming a trend. However, the large amount of data transmission and the computing requirements of intelligent tasks lead to the complex resource management problems. In view of the above challenges, this paper proposes a tasks-oriented joint resource allocation scheme (TOJRAS) in the scenario of IoV. First, this paper proposes a system model with sensing, communication, and computing integration for multiple intelligent tasks with different requirements in the IoV. Secondly, joint resource allocation problems for real-time tasks and delay-tolerant tasks in the IoV are constructed respectively, including communication, computing and caching resources. Thirdly, a distributed deep Q-network (DDQN) based algorithm is proposed to solve the optimization problems, and the convergence and complexity of the algorithm are discussed. Finally, the experimental results based on real data sets verify the performance advantages of the proposed resource allocation scheme, compared to the existing ones. The exploration efficiency of our proposed DDQN-based algorithm is improved by at least about 5%, and our proposed resource allocation scheme improves the mAP performance by about 0.15 under resource constraints.
The growing demand for low delay vehicular content has put tremendous strain on the backbone network. As a promising alternative, cooperative content caching among different cache nodes can reduce content access delay. However, heterogeneous cache nodes have different communication modes and limited caching capacities. In addition, the high mobility of vehicles renders the more complicated caching environment. Therefore, performing efficient cooperative caching becomes a key issue. In this paper, we propose a cross-tier cooperative caching architecture for all contents, which allows the distributed cache nodes to cooperate. Then, we devise the communication link and content caching model to facilitate timely content delivery. Aiming at minimizing transmission delay and cache cost, an optimization problem is formulated. Furthermore, we use a multi-agent deep reinforcement learning (MADRL) approach to model the decision-making process for caching among heterogeneous cache nodes, where each agent interacts with the environment collectively, receives observations yet a common reward, and learns its own optimal policy. Extensive simulations validate that the MADRL approach can enhance hit ratio while reducing transmission delay and cache cost.
The connected autonomous vehicle is considered an effective way to improve transport safety and efficiency. To overcome the limited sensing and computing capabilities of individual vehicles, we design a digital twin assisted decision-making framework for Internet of Vehicles, by leveraging the integration of communication, sensing and computing. In this framework, the digital twin entities residing on edge can effectively communicate and cooperate with each other to plan sub-targets for their respective vehicles, while the vehicles only need to achieve the sub-targets by generating a sequence of atomic actions. Furthermore, we propose a hierarchical multi-agent reinforcement learning approach to implement the framework, which can be trained in an end-to-end way. In the proposed approach, the communication interval of digital twin entities could adapt to time-varying environment. Extensive experiments on driving decision-making have been performed in traffic junction scenarios of different difficulties. The experimental results show that the proposed approach can largely improve collaboration efficiency while reducing communication overhead.
Data sharing technology in Internet of Vehicles(IoV) has attracted great research interest with the goal of realizing intelligent transportation and traffic management. Meanwhile, the main concerns have been raised about the security and privacy of vehicle data. The mobility and real-time characteristics of vehicle data make data sharing more difficult in IoV. The emergence of blockchain and federated learning brings new directions. In this paper, a data-sharing model that combines blockchain and federated learning is proposed to solve the security and privacy problems of data sharing in IoV. First, we use federated learning to share data instead of exposing actual data and propose an adaptive differential privacy scheme to further balance the privacy and availability of data. Then, we integrate the verification scheme into the consensus process, so that the consensus computation can filter out low-quality models. Experimental data shows that our data-sharing model can better balance the relationship between data availability and privacy, and also has enhanced security.
