Vehicular edge computing (VEC) can learn and cache most popular contents for vehicular users (VUs) in the roadside units (RSUs) to support real-time vehicular applications. Federated learning (FL) can protect VUs' privacy by sharing vehicles' local models instead of data. In traditional FL, the global model is periodically updated by aggregating all vehicles' local models. However, vehicles may frequently been driven out of the coverage area of VEC before they finish the local model training and thus the traditional FL cannot upload all local models as expected, which would degrade the accuracy of global model. The asynchronous FL can be performed without aggregating all vehicles' local models, thus more local models can be uploaded to improve the accuracy of global model. The vehicle mobility significantly impacts the asynchronous FL. There is no published work considering the vehicle mobility to design the cooperative caching in VEC based on asynchronous FL. In addition, the caching capacity of RSU is limited and the size of the predicted popular contents usually exceeds the cache capacity of RSU. Hence, VEC should cache the predicted popular contents in different RSUs while considering content transmission delay. In this paper, we consider vehicle mobility and propose a cooperative caching scheme in the VEC based on asynchronous federated and deep reinforcement learning (CAFR). We first consider vehicle mobility and propose an asynchronous FL algorithm to obtain an accurate global model, and then propose an algorithm to predict popular contents based on the global model. In addition, we consider vehicle mobility and propose a deep reinforcement learning (DRL) algorithm to obtain the optimal cooperative caching location for the predicted popular contents in order to optimize the content transmission delay. Extensive experimental results have demonstrated that CAFR scheme outperforms other baseline caching schemes.

Vehicular edge computing (VEC) is envisioned as a promising approach to process the explosive computation tasks of vehicular user (VU). In the VEC system, each VU allocates power to process partial tasks through offloading and the remaining tasks through local execution. During the offloading, each VU adopts the multi-input multi-output and non-orthogonal multiple access (MIMO-NOMA) channel to improve the channel spectrum efficiency and capacity. However, the channel condition is uncertain due to the channel interference among VUs caused by the MIMO-NOMA channel and the time-varying path loss caused by the mobility of each VU. In addition, the task arrival of each VU is stochastic in the real world. The stochastic task arrival and uncertain channel condition affect greatly on the power consumption and latency of tasks for each VU. It is critical to design an optimal power allocation scheme considering the stochastic task arrival and channel variation to optimize the long-term reward, including the power consumption and latency in the MIMO-NOMA VEC. Different from the traditional centralized deep reinforcement learning (DRL)-based scheme, this article constructs a decentralized DRL framework to formulate the power allocation optimization problem, where the local observations are selected as the state. The deep deterministic policy gradient (DDPG) algorithm is adopted to learn the optimal power allocation scheme based on the decentralized DRL framework. Simulation results demonstrate that our proposed power allocation scheme outperforms the existing schemes.

In the rapidly evolving landscape of Internet of Vehicles (IoV) technology, cellular vehicle-to-everything (C-V2X) communication has attracted much attention due to its superior performance in coverage, latency, and throughput. Resource allocation within C-V2X is crucial for ensuring the transmission of safety information and meeting the stringent requirements for ultralow latency and high reliability in vehicle-to-vehicle (V2V) communication. This article proposes a method that integrates graph neural networks (GNNs) with deep reinforcement learning (DRL) to address this challenge. By constructing a dynamic graph with communication links as nodes and employing the graph sample and aggregation (GraphSAGE) model to adapt to changes in graph structure, the model aims to ensure a high success rate for V2V communication while minimizing interference on vehicle-to-infrastructure (V2I) links, thereby ensuring the successful transmission of V2V link information and maintaining high transmission rates for V2I links. The proposed method retains the global feature learning capabilities of GNN and supports distributed network deployment, allowing vehicles to extract low-dimensional features that include structural information from the graph network based on local observations and to make independent resource allocation decisions. Simulation results indicate that the introduction of GNN, with a modest increase in computational load, effectively enhances the decision-making quality of agents, demonstrating superiority to other methods. This study not only provides a theoretically efficient resource allocation strategy for V2V and V2I communications but also paves a new technical path for resource management in practical IoV environments.

This paper studies performance modeling of the IEEE 802.11p enhanced distributed channel access (EDCA) mechanism and develops performance models to analyze the access performance of the IEEE 802.11p EDCA mechanism. A 2-D Markov chain is first constructed to model the backoff procedure of an access category (AC) queue and establish a relationship between the transmission probability and collision probability of the AC queue. Then, a 1-D discrete-time Markov chain is constructed to model the contention period of an AC queue and establish another relationship between the transmission probability and collision probability of the AC queue. Unlike most existing work, the 1-D Markov chain is extended to be infinite in modeling the contention period of an AC queue under both saturated and nonsaturated conditions. The two Markov models take into account all major factors that could affect the access performance of the IEEE 802.11p EDCA mechanism, including the saturation condition, standard parameters, backoff counter freezing, and internal collision. Based on the two Markov models, performance models are further derived to describe the relationships between the parameters of an AC queue and the access performance of the AC queue in terms of the transmission probability, collision probability, normalized throughput, and average access delay, respectively. The effectiveness of the performance models are verified through simulation results.

Vehicular edge computing (VEC) is a promising technology to support real-time vehicular applications, where vehicles offload intensive computation tasks to the nearby VEC server for processing. However, the traditional VEC that relies on single communication technology cannot well meet the communication requirement for task offloading, thus the heterogeneous VEC integrating the advantages of dedicated short-range communications (DSRC), millimeter-wave (mmWave) and cellular-based vehicle to infrastructure (C-V2I) is introduced to enhance the communication capacity. The communication resource allocation and computation resource allocation may significantly impact on the ultra-reliable low-latency communication (URLLC) performance and the VEC system utility, in this case, how to do the resource allocations is becoming necessary. In this paper, we consider a heterogeneous VEC with multiple communication technologies and various types of tasks, and propose an effective resource allocation policy to minimize the system utility while satisfying the URLLC requirement. We first formulate an optimization problem to minimize the system utility under the URLLC constraint which modeled by the moment generating function (MGF)-based stochastic network calculus (SNC), then we present a Lyapunov-guided deep reinforcement learning (DRL) method to convert and solve the optimization problem. Extensive simulation experiments illustrate that the proposed resource allocation approach is effective.

Federated Learning (FL) can protect the privacy of the vehicles in vehicle edge computing (VEC) to a certain extent through sharing the gradients of vehicles’ local models instead of local data. The gradients of vehicles’ local models are usually large for the vehicular artificial intelligence (AI) applications, thus transmitting such large gradients would cause large per-round latency. Gradient quantization has been proposed as one effective approach to reduce the per-round latency in FL enabled VEC through compressing gradients and reducing the number of bits, i.e., the quantization level, to transmit gradients. The selection of quantization level and thresholds determines the quantization error, which further affects the model accuracy and training time. To do so, the total training time and quantization error (QE) become two key metrics for the FL enabled VEC. It is critical to jointly optimize the total training time and QE for the FL enabled VEC. However, the time-varying channel condition causes more challenges to solve this problem. In this paper, we propose a distributed deep reinforcement learning (DRL)-based quantization level allocation scheme to optimize the long-term reward in terms of the total training time and QE. Extensive simulations identify the optimal weighted factors between the total training time and QE, and demonstrate the feasibility and effectiveness of the proposed scheme.