16 August 2024 in IEEE Transactions on Vehicular Technology
It is a great challenge to guarantee both path-tracking performance and vehicle stability when suffering from aggressive uncertainties and severe disturbances. We design a novel learning-based robust model predictive path-tracking controller to alleviate the influence of disturbances, avoid over-conservative steering actions, and mediate the conflict between path-tracking and vehicle stability. Specifically, we firstly utilize the friction limits of tires and define an enveloped stable zone in the phase portrait which is used as safety constraints. Secondly, two approaches under the model predictive control (MPC) framework are employed to tackle the severe uncertainties and disturbances: 1) a deep neural network (DNN) dynamics model is employed to estimate the predictive error and attenuate the mismatch between the nominal model and actual plant; and 2) a local feedback linear quadratic regulator (LQR) is used to stabilize the system matrix, calculate invariant tube, and thus guarantee all state and control constraints are satisfied. Finally, real vehicle experiments indicate that the proposed controller can achieve an over 18% improvement in path-tracking performance and guarantee vehicle stability, even for cases with severe uncertainties and disturbances.
22 November 2024 in Advanced Engineering Informatics
25 December 2024 in Engineering Applications of Artificial Intelligence
12 March 2026 in IEEE Transactions on Artificial Intelligence
The motion prediction of surrounding vehicles is crucial for autonomous driving. As it can effectively guide the planning and decision-making of the autonomous vehicle. Although many methods have been proposed to predict the future motion of vehicles, but they failed capture the essense of multitask learning for vehicle motion prediction and lacked effective semantic and interaction information fusion. In this paper, we propose a novel model, Graph Attention Transformer with Multi- Task Learning (GATran), to predict the future motion of the target vehicle by modeling motion prediction task as a multi-task learning problem of trajectory prediction and driving intention estimation. The model consists of three modules: Feature Fusion, Spatio-Temporal Interaction, and Motion Decoder. The Feature Fusion module is used to effectively fuse the semantic and dynamic information of the target vehicle. The Spatial Interaction module is used to model the interaction between vehicles. The Motion Decoder module decouples the motion prediction into multi-task learning for trajectory prediction and driving intention estimation in a parallel manner. Extensive experiments on real-world datasets, with average improvement of 50% in NGSIM and 40% in HighD, demonstrate that the proposed model achieves superior performance.