A major obstacle to the rapid maturation and real-world deployment of cooperative connected and automated vehicles (CAVs) is the prohibitive cost and extensive on-road testing mileage required to validate safety in natural traffic, where genuinely high-risk scenarios are exceedingly rare. Although existing scenario-generation methods can generate high-risk scenarios at scale, such scenarios frequently violate real-world physics or driver-behavior patterns, making them implausible and unsuitable for rigorous evaluation. To bridge this critical gap, we propose a plausible high-risk scenario generation method utilizing a bidirectional and autoregressive transformer (BART). Continuous vehicle trajectories from extensive naturalistic datasets are tokenized into a concise behavioral vocabulary, enabling the model to capture latent plausibility structures and realistically reproduce vehicle maneuvers. An iterative risk-feedback mechanism further steers scenario generation toward aggressive yet physically plausible driving conditions, effectively escalating cumulative risk within each simulation and thus yielding more plausible high-risk scenarios. Across cooperative lane-change and merging verification, the proposed BART-driven plausible high-risk generator yields markedly more high-risk and informative test scenarios than the Markov Decision Process (MDP) baseline, an Adaptive Stress Testing with a Deep Q-Network (AST-DQN), and a BART variant without risk-guided decoding, while maintaining a practical balance between scenario plausibility and risk elevation.

Highway interchanges are critical nodes in transportation networks but frequently experience congestion due to complex interactions between ramp diverging and merging flows. While Connected and Automated Vehicles (CAVs) are expected to improve traffic efficiency, the coexistence of Human-Driven Vehicles (HDVs) introduces disturbances that undermine performance. Prior studies have addressed certain cooperative strategies of CAVs in mixed traffic, but they often overlook the interactions between adjacent bottlenecks in interchange scenarios and the heterogeneity of HDV driving styles. To address this gap, this study proposes an Idealization-Oriented Planning (IOP) scheme that enhances interchange traffic efficiency in mixed traffic with high CAV penetration. Enabled by Cloud Control Systems (CCS), an idealized optimum for the overall travel efficiency under the assumption of full CAV penetration is firstly derived as a reference, based on which short-horizon strategies are computed to mitigate disturbances caused by HDV behaviors. Specifically, a DeePC-based Lane-Change Hesitation Guidance mechanism identifies conservative HDVs in advance and adjusts lane-change gaps to prevent excessive deceleration. In parallel, an Aggressive Merging Avoidance mechanism formulates a potential game in which adjacent CAVs cooperate to constrain inefficient HDV cut-ins, yielding safe and system-efficient strategies. A Python-based simulation platform validates the proposed approach, showing that IOP outperforms benchmark methods across various traffic conditions and HDV penetration rates (5%–30%); as an illustrative case, under high-flow conditions with 15% HDV penetration, IOP achieves a delay reduction of up to 53%.

Cooperative lane-change is a pivotal application of connected and automated vehicles (CAVs), enhancing traffic safety and efficiency, especially in congested urban areas and highway on-ramps. However, the widespread implementation of multi-CAV cooperative lane-change is hindered by the lack of efficient verification strategies. This issue is exacerbated by the “curse of dimensionality”, stemming from numerous scenario variables and complex algorithms. To overcome this, we propose an efficient verification strategy that accelerates the process through optimal feature selection. By eliminating variables that have minimal impact on evaluation outcomes yet significantly increase testing complexity, our approach streamlines the verification process, minimizing information loss while maintaining high efficiency. We validated this strategy in various urban and highway scenarios involving multiple CAVs. During the verification process, the dimensionality of scenario variables was reduced, resulting in an exponential decrease in the number of testing scenarios, with information loss limited to no more than 6%. The results demonstrate that our proposed strategy significantly improves the efficiency of verifying cooperative lane-change algorithms in high-dimensional scenario variables, with minimal loss of essential information.

The multi-vehicle coordinated lane change is one typical application of intelligent connected vehicle(ICV), which must be systematically and thoroughly verified before across-the-board commercial application. Existing evaluation frameworks face challenges in effectively verifying multi-vehicle coordinated lane change algorithm, whose decision-making process is more complex and needs to consider more complex surrounding environments. This complexity introduces the "curse of dimensionality" into the verification process, adversely impacting verification efficiency. To address the aforementioned challenge, an efficient verification strategy with optimal feature subset selection is proposed in this study. Initially, the subset feature is defined by the integrals of position probability density function between host vehicle and surrounding vehicles across various decision-making phases of coordinated lane change algorithm. Following this, the optimal feature subset selection method is presented for verification in different decision-making phases of coordinated lane change algorithm. Subsequently, the verification strategy is delineated. Finally, the optimal feature subset selection verification strategy is implemented within a coordinated lane change scenario. A multi-start search algorithm is employed to explore the feasible domain of the multi-vehicle coordinated lane change algorithm. Verification through simulation is then executed, and its efficiency is compared with a widely used evaluation framework based on Test Matrix. Notably, the proposed strategy demonstrates a minimum efficiency improvement of 85%. The verification results underscore the effectiveness the proposed method in verification of phased multi-vehicle coordinated lane change decision-making algorithm, particularly within high-dimensional and complex environments.

The on-ramp merging is one of the typical bottlenecks on highways, and it’s expected to improve vehicle safety and traffic efficiency in this area through multi-vehicle collaboration. Existing research rarely coordinates on-ramp merging utilizing global information in a cyber-physical system, and most of them assume that vehicles in the mainline wouldn’t change lanes for simplification. However, scheduling methods dealing with multi-lane merging areas have been less explored. To address the problem, an optimized scheduling method with dynamic conflict graph is proposed in this study. First, the dynamic conflict graph is established, where vertices define the attributes of vehicle groups and edges describe the relationship among them; the optimization problem is then reconstructed as a graph search problem. Subsequently, a graph decomposition method is presented for the dynamic conflict graph. The feasible domain of vertices’ final states and costs of edges are determined based on optimal control theory, after which the heuristic depth-first search strategy is adopted to find a near-optimal solution. Finally, the dynamic conflict graph is applied in a continuous traffic flow. Simulations are conducted, and the performance is compared with the default algorithm in SUMO. The simulation results reveal that the proposed method reduces the overall travel delay while guaranteeing safety.

Cut-in critical level prediction is of vital importance to Advanced Driving Assistance system (ADAS) to fulfil regional requirements and to increase safety. Time-to-Collision is the key component for critical levels of cut-in scenario. Hence, a new simulation based Time-to-Collision (TTC) calculation algorithm is firstly introduced in this paper. For the purpose of cut-in critical level prediction, the values of TTC in c.a 2000 cut-in cases are calculated, which are used to train a novel machine learning based cut-in critical level prediction method. The goal of ADAS functions development is to perform as a sophisticated driver, especially in dealing with risks. Thus, the correlation coefficient between ego vehicle deceleration and TTC could be used to evaluate the performance of different TTC calculation methods. In order to validate the superiority of simulation based TTC calculation algorithm, the Pearson correlation coefficient is calculated for the simulation based TTC and the TTC calculated by the traditional method, which are 0.7882 and 0.1357, respectively. Through enough valid regional cut-in samples trained prediction algorithm, TTC could be estimated accurately and effectively, i.e., the accuracy reaches 92%. To the best of the author's knowledge the simulation based TTC calculation method and the cut-in critical level prediction learning algorithm are new contributions in ADAS field.