Robust and reliable human–machine shared driving (HMSD) is essential for balancing safety and comfort. Within a connected urban arterial system, rare high-risk long-tail disturbances can trigger conflicts, lane departures, and oscillatory flow, degrading system safety and efficiency of the HMSD. To mitigate these effects, an interactive learning framework is built by coupling a scalable environment model with a dual-discriminator generative adversarial network to synthesise diverse, high-fidelity extreme scenarios, thereby enlarging the training domain of HMSD. A bidirectional loop between the environment model and the decision maker enables continuous refinement of the control policy, risk suppression, and improvement in response efficiency. The trained controller leverages cooperative vehicle–infrastructure sensing to derive shared risk states and to adaptively allocate authority between the human driver and automation in real time. The robustness of the proposed method is validated by comparing it with other shared control frameworks on a hierarchical validation platform, including a driver-in-the-loop (DiL) system. The results demonstrate that this method offers a superior balance between driving safety, stability, and pleasure, while demonstrating practical and robust control performance under long-tail events.

The learning efficiency remains a critical impediment to the practical application of connected and automated vehicles (CAVs). This paper proposes a transferable model-based reinforcement learning (TMBRL) strategy to enhance the sample efficiency and learning rate of CAVs. Specifically, a surrogate policy model is established by capturing state transitions between the actual environment and the vehicle within traffic scenarios. Then, a model-based reinforcement learning (MBRL) approach is established utilizing a surrogate model and a soft actor-critic algorithm. To improve the learning efficiency of platoon control algorithm, a transfer learning method is implemented to MBRL framework. Specifically, the trained surrogate model of vehicles in the source domain is transferred to vehicles in the target domain, and the latter just should update the surrogate model in terms of the individual dynamic characteristics and tasks. Finally, a platoon experiment platform with Prescan software is conducted. The experimental evaluation demonstrates that the TMBRL strategy significantly outperforms conventional reinforcement learning approaches, achieving higher average cumulative reward of 47 and demonstrating a 16% improvement in training success rate. Comparative analysis further reveals that the proposed TMBRL strategy exhibits superior robustness in platoon tracking tasks, maintaining enhanced trajectory tracking precision and stability under dynamic environmental conditions.

Integrating range-extended electric vehicles (REEVs) in the automotive market is a key part of the drive toward environmental sustainability. This paper leverages an experience-shared approach to variable-step predictive control to improve REEV energy efficiency, where a transferable driver model is designed to accommodate varying driver experience levels via knowledge transfer. This model incorporates a confidence level factor to determine the effective length of speed prediction, ensuring a more accurate and reliable model predictive control system with lower requirement data. A grey wolf optimizer is employed as an advanced global solver in the model predictive control system of the studied REEV to seek better energy-saving performance. Experimental validation utilizes an industry-recognized driver-in-the-loop co-simulation platform to investigate the proposed approach’s performance. Compared to Gaussian mixture regression one, the transferable driver model achieves a 27.29% improvement in speed prediction accuracy. Incorporating the driver model, the proposed experience-shared variable-step predictive control approach helps a 3.9% reduction in fuel consumption compared to an LQR-driven MPC one.

This paper proposes a driver-centric data-driven model predictive control (DDMPC) strategy to improve driving comfort while maintaining driving safety of vehicular platoon. This strategy combines a data-driven model predictive controller and the driver-centric driving policy. The data-driven platoon model involving longitudinal-lateral dynamics is established with subspace identification to alleviate the adverse effects of uncertain dynamics. Then, a subspace predictor-based distributed data-driven model predictive controller is developed for vehicular platoon. To overcome the cutting-corner phenomenon on curved roads, the reference point is shifted from the preceding vehicle to an optimal corridor point behind it. In this way, a driver-centric driving policy is designed with a flexible spacing and soft control constraints to balance driving safety and driving comfort in terms of different driving styles. Finally, several experiments with sixty drivers are carried out on a self-developed vehicular platoon platform. The experimental results demonstrate the effectiveness of the proposed DDMPC strategy.

Dedicated hybrid engine technology using auxiliary electronic components has been proven as an energy-saving solution to public concerns about energy consumption and carbon emissions. This paper proposes a generic approach of K-fold filter-collaborated feature selection (KFFC-FS) to robustly identify the key parameters of three engine performance indicators, i.e., volumetric efficiency, thermal efficiency, and fuel consumption. By using this approach, five filters are collaborated to provide a robust rank of feature importance and avoid the feature overestimation caused by the single filter. Meanwhile, the K-fold cross validation method is introduced to avoid random precision issues and overfitting, further enhancing the robustness of key parameter identification for the independent engine performance indicators. In this research, the modelling data is collected from an experimental test bench with a BYD 1.5L gasoline engine. Under the basics of the studied three engine performance indicators by using a multiple-layer perceptron network, the proposed approach further reduces by at least 10.3% root-mean-square error (RMSE) and at least 30% reduction of the model inputs.

The accurate prediction of soft sensors is essential for the development of modern combustion engines to achieve better performance, lower emissions, and reduced fuel consumption. To precisely predict engine performance, i.e., indicated thermal efficiency, volumetric efficiency, and fuel consumption rate of a hybrid engine, in this article, we propose a novel data-driven approach of statistics-guided accelerated swarm feature selection to find the most effective features for engine soft sensors. Differing from the existing filter or wrapper feature selection approaches, this approach uses external measure information to direct velocity updates in the accelerated swarm feature selection. Several filter and wrapper methods are developed and comprehensively compared. The experimental dataset is collected from a BYD 1.5 L gasoline engine. Validated by bench test, the results demonstrate that the proposed approach finds the most effective features and optimal network structure for data-driven performance prediction of the hybrid engine that was studied.

This article proposes a new form of algorithm environment for the multiobjective optimization of an energy management system in plug-in hybrid vehicles (PHEVs). The surrogate-assisted strength Pareto evolutionary algorithm (SSPEA) is developed to optimize the power-split control parameters guided by the data from the physical PHEV and its digital twins (DTs). By introducing a “confidence factor,” the SSPEA uses the fused data of physically measured and virtually simulated vehicle performances (energy consumption and remaining battery state of charge) to converge the optimization process. Gaussian noisy models are adopted to emulate the real vehicle system on the hardware-in-the-loop platform for experimental evaluation. The testing results suggest that the proposed SSPEA requires less R&D costs than the model-free method that only uses the physical information, and more than 44.6% energy can be saved during the R&D process. Driven by the SSPEA, the optimized energy management system surpasses other non-DT-assisted systems by saving more than 4.8% energy.

This article introduces the concept of the driver-identified supervisory control system, which forms a novel architecture of adaptive energy management for hybrid electric vehicles (HEVs). As a man-machine system, the proposed system can accurately identify the human driver from natural operating signals and provides driver-identified globally optimal control policies as opposed to mere control actions. To help improve the identifiability and efficiency of this control system, the method of spectrum-guided fuzzy feature extraction (SFFE) is developed. First, the configuration of the HEV model and its control system are analyzed. Second, design procedures of the SFFE algorithm are set out to extract 15 groups of features from primitive operating signals. Third, long-term and short-term memory networks are developed as a driver recognizer and tested by the features. The driver identity maps to corresponding control policies optimized by dynamic programming. Finally, the comparative study includes involved extraction methods and their identification system performance as well as their application to HEV systems. The results demonstrate that with help of the SFFE, the driver recognizer improves identifiability by at least 10% compared to that obtained using other involved extraction methods. The improved HEV system is a significant advance over the 5.53% reduction on fuel consumption obtained by the fuzzy-logic-based system.

This paper researches on a cyber-physical energy-saving control framework for a plug-in hybrid aircraft-towing tractor, in which, an online optimization methodology named the online swarm intelligent programming (OSIP) is proposed. The new methodology obtains real-time optimal control signals from the vehicle to everything (V2X) network, and the widely used charge depleting/charge sustaining strategy is upgraded to a more adaptive and intelligent level. The energy flow of the hybrid aircraft-towing tractor with connectivity is first analyzed and modeled for OSIP. The optimal control problem is then formulated as an online integer optimization and the OSIP algorithm based on chaos-enhanced accelerated swarm optimization is developed to minimize the powertrain power loss in real-time. Finally, the advantages of the new energy management system are demonstrated and evaluated by hardware-in-the-loop testing. The results show that up to 17% fuel and 13% total energy loss can be saved via the proposed cyber-physical control.