Traffic speed prediction based on real-world traffic data is a classical problem in intelligent transportation systems (ITS). Most existing traffic speed prediction models are proposed based on the hypothesis that traffic data are complete or have rare missing values. However, such data collected in real-world scenarios are often incomplete due to various human and natural factors. Although this problem can be solved by first estimating the missing values with an imputation model and then applying a prediction model, the former potentially breaks critical latent features and further leads to the error accumulation issues. To tackle this problem, we propose a graph-based spatio-temporal autoencoder that follows an encoder-decoder structure for spatio-temporal traffic speed prediction with missing values. Specifically, we regard the imputation and prediction as two parallel tasks and train them sequentially to eliminate the negative impact of imputation on raw data for prediction and accelerate the model training process. Furthermore, we utilize graph convolutional layers with a self-adaptive adjacency matrix for spatial dependencies modeling and apply gated recurrent units for temporal learning. To evaluate the proposed model, we conduct comprehensive case studies on two real-world traffic datasets with two different missing patterns and a wide and practical missing rate range from 20% to 80%. Experimental results demonstrate that the model consistently outperforms the state-of-the-art traffic prediction with missing values methods and achieves steady performance in the investigated missing scenarios and prediction horizons.

Electric vehicle (EV) dynamic wireless charging system has become an emerging application in the area of the intelligent transportation system (ITS). However, an integrated design of the EV dynamic wireless charging system requires the considerations of both the economical and technical perspectives of a smart city. Specifically, most of the existing researches unilaterally consider the application of either placement strategy for power tracks (PTs) or dynamic vehicle-to-grid (V2G) scheduling. In this article, we propose a multistage system framework to account for an integrated EV dynamic wireless charging system in a smart city. First of all, an optimal placement strategy for PTs is developed based on city traffic information and EV energy demand. Then, having the optimal locations of PTs through the previous stage approach, the proposed dynamic V2G scheduling scheme is formulated to coordinate the schedules of EVs with the provision of daytime V2G ancillary services. Our simulation results present that the proposed multistage system model achieves improvements on both the placement strategy and V2G scheduling scheme. In addition, relatively low economic system costs can be obtained through our proposed model.

Rate-Splitting Multiple Access (RSMA) has recently found favour in the multi-antenna-aided wireless downlink, as a benefit of relaxing the accuracy of Channel State Information at the Transmitter (CSIT), while in achieving high spectral efficiency and providing security guarantees. These benefits are particularly important in high-velocity vehicular platoons since their high Doppler affects the estimation accuracy of the CSIT. To tackle this challenge, we propose an RSMA-based Internet of Vehicles (IoV) solution that jointly considers platoon control and FEderated Edge Learning (FEEL) in the downlink. Specifically, the proposed framework is designed for transmitting the unicast control messages within the IoV platoon, as well as for privacy-preserving FEEL-aided downlink Non-Orthogonal Unicasting and Multicasting (NOUM). Given this sophisticated framework, a multi-objective optimization problem is formulated to minimize both the latency of the FEEL downlink and the deviation of the vehicles within the platoon. To efficiently solve this problem, a Block Coordinate Descent (BCD) framework is developed for decoupling the main multi-objective problem into two sub-problems. Then, for solving these non-convex sub-problems, a Successive Convex Approximation (SCA) and Model Predictive Control (MPC) method is developed for solving the FEEL-based downlink problem and platoon control problem, respectively. Our simulation results show that the proposed RSMA-based IoV system outperforms both the popular Multi-User Linear Precoding (MU–LP) and the conventional Non-Orthogonal Multiple Access (NOMA) system. Finally, the BCD framework is shown to generate near-optimal solutions at reduced complexity.

Autonomous vehicles (AVs), as one of the cores in future intelligent transportation systems (ITSs), can facilitate reliable and safe traffic operations and services. The ability to automatically perform effective AV motion planning and deploy efficient perception systems is vital for advancing the quality of core transportation services. However, existing research studies have only considered the applications of either of these approaches, which neglect their necessary interactions in real-world AV motion planning systems. To address this problem, we design an AV motion planning strategy based on motion prediction and V2V communication. Specifically, we propose the perception system and V2V communication module to provide real-time traffic and vehicular information to the participated AVs. Then, we formulate the AV lane-change motion planning problem through the scope of model predictive control based problem, as well as proposing the method on learning optimal motion planning by means of a novel deep learning technique. We conduct extensive case studies to evaluate the performance of the proposed system model. Our experimental results demonstrate the effectiveness of the proposed system model under various traffic conditions. In addition, the robustness of the perception system is guaranteed by utilizing the Car Learning to Act (CARLA) system with available V2V communication.

The growth of vehicular applications will inevitably require Base Stations (BSs) to simultaneously serve more Connected Vehicles (CVs) within limited bandwidth resources, which imposes a great challenge in interference management. Effective management of this interference is crucial for reliable Vehicle-to-Everything (V2X) communication, and necessitates accurate Channel State Information at the Transmitter (CSIT). In practice, the dynamic and unpredictable nature of CV movements prevents BS from obtaining perfect CSIT, leading to outdated information and threatening communication performance. In this study, we propose a Rate-Splitting Multiple Access (RSMA)-enabled V2X communication system to efficiently manage interference channels. We leverage a 1-layer RSMA scheme to relax the stringent requirement for perfect CSIT and enhance robustness to outdated information. Furthermore, we introduce Gruformer, a transformer-based model for improved CSIT prediction utilizing historical data. While longer forecasting horizons decrease accuracy, we present a game theory-based approach that significantly reduces processing time for power allocation, enabling timely decisions before CSIT becomes outdated. Simulation results reveal that Gruformer allows for more accurate predictions during rapid changes in channel conditions. Leveraging this high-quality CSIT, the proposed V2X system achieves a 20% increase in Weighted Ergodic Sum-Rate (WESR). Furthermore, the game theory-based approach delivers a 60% reduction in processing time while maintaining near-optimal performance.

In a smart grid integrated with vehicle-to-grid (V2G) technique, electric vehicles (EVs) fleets under effective coordination can be considered as a massive aggregated power storage to provide frequency regulation service. In this article, we propose a hierarchical system model to jointly optimize power flow routing and V2G scheduling for providing regulation service. First of all, by installing power flow routers (PFRs) inside the power grid, we formulate the problem of optimal power flow (OPF) routing at the grid level. Through the utilization of the semidefinite programming (SDP) relaxation, we can transform the original non-deterministic polynomial-time hard (NP-hard) problem into a convex problem. The tree decomposition method is then used to further reduce the complexity of the system network. After solving the grid-level OPF routing problem, a forecast-based scheduling problem is formulated at the EV level to coordinate EVs by providing the V2G regulation service. To cope with the forecast uncertainties, an online scheduling problem is in turn formulated. In order to solve these problems in a scalable manner, decentralized algorithms are then devised to control the EV schedules. The simulation results show that the devised online scheduling algorithm can outperform the existing algorithms, which is able to flatten the power fluctuations at the buses with EVs attached. Additionally, we show that grid stability issue can be alleviated through the proposed model. Finally, for different power systems, the uses of PFRs can reduce the system power loss in a great manner while providing voltage regulation.