Autonomous driving often suffer from a decoupled feedback loop between prediction and planning. While prediction losses focus on the accuracy of surrounding agents, planning losses typically imitate recorded Autonomous Vehicles’ (AVs) trajectories that may contain suboptimal or aggressive behaviors, leading to unstable interactions in mixed traffic. This paper presents HybridLoss, an adaptive planning-oriented objective that unifies prediction and planning through planner-in-the-loop supervision and interaction-aware consistency. HybridLoss integrates an adaptive motion-planning module which replaces ground-truth targets with optimized reference trajectories, and a multi-term loss combining prediction, adaptive planning, safety potential, and social force objectives. Evaluations on the INTERACTION dataset indicate that HybridLoss significantly outperforms strong baselines. Beyond standard metric improvements—reducing ADE/FDE from 1.36/1.64 m to 1.11/1.36 m and collision rates from 0.19% to 0.11%—extensive stress-testing reveals superior system maturity. First, HybridLoss exhibits the highest robustness under input perturbations, maintaining the lowest planning deviation and endpoint standard deviation. Second, it demonstrates strong generalization, maintaining stable success rates (87.7%) in unseen scenarios with high computational efficiency (64.3 Hz). Third, multi-objective analysis confirms that HybridLoss achieves the optimal Pareto trade-off between efficiency, safety, and comfort, avoiding the speed-safety collapse seen in baseline methods. Finally, social force evaluations highlight that HybridLoss fosters implicit cooperation, achieving higher yield rates and reduced conflict indices while maintaining safe interaction buffers. These results validate HybridLoss as a robust, socially compliant, and adaptive solution for end-to-end driving.

Errors are ubiquitous under the influence of multifarious uncertain factors in real traffic, causing information that used to regulate the mobility of connected automated vehicles (CAVs) to be inconsistent with true information, which provoke concerns on CAVs platoon performance. To date, due to the lack of appropriate model and analytical methodology, our understanding of influence mechanisms of various information errors on CAVs and the resulting anomalous platoon dynamics remains limited. Therefore, this study develops a comprehensive framework to fill these gaps. First, we expound the possible information errors in CAVs, upon which we categorize them into three types: source error, transmission error, and latency error. Second, we formulate a generic bi-layer architecture model with dynamical information flow topology, incorporating the information with errors from both of vehicular and communication layers, so as to describe the longitudinal dynamics of CAVs with information errors. Third, we derive the closed-loop and pseudo string stability conditions of CAVs platoon in the presence of information errors. Final, we perform extensive numerical simulations to exposit the effects of three types of information errors on CAVs platoon dynamics. The results show that information errors may cause deteriorated platoon stability, oscillations, reduced traffic efficiency, wasted road resources, and even collision accidents, which undoubtedly degrade the platoon performance. Overall, this study uncovers how different types of information errors influence the individual behaviors of CAVs and thus the longitudinal dynamics evolution of platoon, laying foundation to establish corresponding anomaly detection and counteracting control methods.

This study investigates the dispatcher number (DN) of one-way carsharing services (CSSs), considering the critical yet overlooked aspect of dispatchers’ break requirement with vehicle relocation and dispatcher assignments. The DN problem aims to minimize the daily cost of one-way CSS operators by determining the fleet size, the number of dispatchers, vehicle relocation, and dispatcher movement under the restriction on the maximum accumulative working time of dispatchers. The novelty of the study lies in the incorporation of the break requirement consideration of dispatchers into the personnel assignment during the operation period. A nonlinear integer programming (NLIP) model is first developed for the DN problem. By exploring the structure of the proposed model, an effective heuristic solution method consisting of optimization and simulation modules is proposed to obtain the optimal solution to the problem. Numerical experiments based on EVCARD, a popular one-way CSS operator in China, are conducted to demonstrate the effectiveness of the proposed models and solution method and the rationality of incorporating the break requirement into the decision-making process. Furthermore, we investigated the impacts of critical parameters, such as payment for dispatchers and relocation cost, on the performance of the one-way CSS, providing valuable insights for carsharing service operators.

Maximum platoon size (MPS) playing a crucial role in the configuration of connected and automated vehicle (CAV) platoons can significantly affect the traffic operation performance at intersections. The study addresses the MPS for platoon-based cooperative signal-free intersection control (PCSIC) problem considering the platoon formation process of the CAVs, which manage the CAVs to form the CAV platoons under the tactical platoon size limit to pass the intersection cooperatively. A mixed-integer nonlinear programming model is first developed to solve the proposed problem by minimizing the total travel delays of all CAVs while guaranteeing the feasible trajectories of the CAVs for platoon formation. A hybrid artificial bee colony algorithm integrating the artificial bee colony approach and dynamic programming algorithm for an optimal control scheme is proposed to solve the proposed model. Numerical experiments are conducted to examine the efficacy of the proposed model and solution algorithm. The impacts of balanced and unbalanced motion states between approaching directions on the MPS are investigated, and the traffic throughput performance of the platoon-based control strategy under balanced and unbalanced scenarios is evaluated. The findings can provide useful insights to transport authorities and automotive operators for signal-free intersection management.

Vehicle platoon has the potential to significantly improve traffic throughput and reduce fuel consumption and emissions and thus has attracted extensive attention recently. In this study, we propose a vehicle platoon of connected and automated vehicles (CAVs) with a combined spacing policy to enhance traffic performance. First, a combined spacing policy composed of the constant time gap (CTG) and constant spacing (CS) is formulated for the proposed vehicle platoon, where the leader adopts the CTG and the followers use the CS policy. Based on the$h_{2}$-norm string stability criteria, the notion of exogenous-head-to-tail string stability is newly introduced, and the sufficient conditions of the local stability and string stability in the frequency domain are derived using the Routh-Hurwitz criterion and Laplace transform respectively. Numerical experiments are conducted to validate the string stability. The effectiveness of the proposed vehicle platoons is verified by theoretical analysis and numerical experiments using two typical scenarios and several measurements of effectiveness (MOE) in various performance aspects, including efficiency, safety, energy, and emission. The results show that the proposed vehicle platoon performs better than the CS-based vehicle platoon in all aspects except for efficiency. It also indicates that the proposed vehicle platoon has obvious advantages over the CTG-based vehicle platoon in efficiency and safety aspects. The findings have demonstrated the merits of the combined application of CTG and CS policies for the vehicle platoon in enhancing stability and traffic performance.