Collision risk modeling with multiple surrounding target vehicles (TVs) is essential for host vehicle (HV) trajectory planning, especially considering challenging TV lateral behaviors. Existing motion-compensated spatial methods ignore TV lateral motion states such as lateral velocity and yaw rate, so that TV lateral behavior cannot be described accurately. Aiming at high-accuracy collision risk modeling, a twisted Gaussian risk model using both longitudinal and lateral motion states for TV behavior description is proposed. Firstly, the HV-TVs system is treated as the superposition of multiple HV-TV units, and a Gaussian risk model is adopted for the collision risk description of the HV-TV unit. Then, by expanding the variances, TV longitudinal and lateral velocities are considered. At last, a twisted Gaussian risk model considering TV yaw rate is constructed based on the projection of the Gaussian risk model. With this twisted Gaussian risk model, TV longitudinal-lateral motion states are considered simultaneously, and TV behaviors can be described for HV-TVs collision risk modeling. For HV trajectory planning, trajectory candidates generated by the maneuver-inspired method are evaluated via the proposed risk model, and the safe and efficient trajectory is selected. Simulation and hardware-in-the-loop experimental results show that the proposed method considering TV longitudinal-lateral motion states allows HV to operate more safely and efficiently than the conventional method.

Reliable trajectory prediction of preceding target vehicles (PTVs) is crucial for the planning and decision making of automated vehicles. However, the future trajectory is affected by the driver's intention and diverse driving styles, which can hardly be predicted precisely, especially when the vehicle performs a lane change maneuver. In this study, we propose a lane crossing and final points generation (CFPG) model-based trajectory prediction approach for PTVs, in which the key influence factors such as the driver's intention and the mixed driving style are included. Firstly, we build a maneuver and stage recognition model upon the long short term memory (LSTM) to infer the current maneuver of the preceding target vehicle. Furthermore, the approach predicts the lane crossing point using a physics-based model combining with a deep conditional generative model trained by a deep neural network. Moreover, a maneuver-based model is adopted to predict the final point according to the prediction interval. In order to avoid the possible cumulative error caused by iteratively generating trajectories in traditional methods, we use a curve fitting method to obtain the predicted trajectory. Lane changing data collected from naturalistic driving scenarios are used to verify the proposed approach, and the results suggest more accurate and reliable prediction trajectories compared with conventional methods.

Autonomous vehicle requires a high-precision lateral dynamics model for path following and lateral stability control. However, existing physical models suffer from low accuracy due to modeling simplification and inaccurate model parameters, while data-driven models lack physical interpretability and robustness. To address these issues, a hybrid architecture inspired by vehicle lateral dynamics is developed in this study, which embeds the data-driven model into a physical model for parameter identification and error characterization to achieve accurate and interpretable modeling. Specifically, a physical lateral dynamics model with error analysis is established at first, and the problems of modeling error characterization and parameter identification are formulated. Then, the physical lateral dynamics model is deformed, and the modeling errors and cornering stiffness are unified into compound parameters. Using this deformed physical model, the modeling errors can be characterized by the identification of these compound parameters. To obtain high-precision compound parameters, a neural network-based parameter identification method is proposed, and the identified time-varying parameters enable high-precision characterization of modeling errors and parameters using data knowledge. By embedding the neural network into the deformed physical model, a hybrid model integrating physical laws and data knowledge is finally established for the description of vehicle lateral dynamics. Simulation and experimental results demonstrate that the proposed hybrid model realizes more accurate modeling of vehicle lateral dynamics than conventional physical and data-driven models.

Accurate trajectory prediction of surrounding vehicles is vital for automated vehicles to achieve high-level driving safety in complex situations. However, most state-of-the-art approaches for multi-vehicle trajectory prediction ignore vehicle motion uncertainty caused by different driving styles. Moreover, the interrelationship between the vehicle and the environment is seldom considered. To address the above problems, this paper proposes a driving risk map-integrated deep learning (DRM-DL) method for interactive trajectory prediction of surrounding vehicles, which comprehensively considers the motion uncertainty, trajectory intention uncertainty and interactions among vehicles, lane lines and road boundaries. Specifically, we adopt a conditional variational autoencoder (CVAE) to generate the candidate trajectories, in which the motion uncertainty is considered using a conditional Gaussian distribution. Furthermore, a driving risk map is constructed to realize a unified and interpretable representation of vehicle-vehicle and vehicle-environment interactions. The probability of each candidate trajectory is assigned using a trajectory probability model and a random selection is adopted to select a guided trajectory, which simulates the driver’s trajectory intention uncertainty. Finally, a relearning module is designed to obtain the precise trajectory prediction for surrounding vehicles. The proposed method is evaluated on the HighD dataset, and the results demonstrate a more accurate and reliable trajectory prediction for surrounding vehicles compared with state-of-the-art methods.

This article addresses the issue of networked lane keeping security control for autonomous vehicles subject to aperiodic controller-targeted denial-of-service (DoS) attacks, taking into account time-varying driving speed and nonlinear tire cornering stiffness. To accurately estimate the incompletely measured states and capture the dynamic behaviors appearing at the end of aperiodic DoS attacks, full state gain adjustable switching observer is established for the fuzzy vehicle-road integrated dynamic systems obtained via the tensor product model transformation method. In order to ensure the control quality and simultaneously save the communication resource, new resilient adaptive event triggered scheme is proposed with reinforcement learning-based intelligent optimal threshold regulation mechanism based on observed states. Then, an augmented observer-based fuzzy switching system is constructed using the time delay method. Additionally, sufficient conditions are established to guarantee global exponential stability of the closed-loop nonlinear lane keeping system with prescribed H∞ performance using piecewise Lyapunov functional analysis approach. Subsequently, the gains for controller, observer and trigger are co-designed and computed by solving certain matrix inequalities. Finally, the effectiveness of proposed security control method is demonstrated through typical maneuver scenario in terms of reasonable triggered times, better tracking performances and acceptable lateral dynamics.

Cooperative perception can significantly improve the perception performance of autonomous vehicles beyond the limited perception ability of individual vehicles by exchanging information with neighbor agents through V2X communication. However, most existing work assume ideal communication among agents, ignoring the significant and common interruption issues caused by imperfect V2X communication, where cooperation agents can not receive cooperative messages successfully and thus fail to achieve cooperative perception, leading to safety risks. To fully reap the benefits of cooperative perception in practice, we propose V2X communication INterruption-aware COoperative Perception (V2X-INCOP), a cooperative perception system robust to communication interruption for V2X communication-aided autonomous driving, which leverages historical cooperation information to recover missing information due to the interruptions and alleviate the impact of the interruption issue. To achieve comprehensive recovery, we design a communication-adaptive multi-scale spatial-temporal prediction model to extract multi-scale spatial-temporal features based on V2X communication conditions and capture the most significant information for the prediction of the missing information. To further improve recovery performance, we adopt a knowledge distillation framework to give explicit and direct supervision to the prediction model and a curriculum learning strategy to stabilize the training of the model. Experiments on three public cooperative perception datasets demonstrate that the proposed method is effective in alleviating the impacts of communication interruption on cooperative perception. V2X-INCOP outperforms state-of-the-art cooperative perception methods and has a cooperative perception gain up to 14.06%, 13.9%, and 12.07% over individual perception on average of different packet drop rates on OPV2V, V2X-Sim, and Dair-V2X datasets, respectively.