The field of autonomous driving and traffic prediction is rapidly advancing, with a focus on improving the accuracy and efficiency of traffic forecasting and vehicle control. Recent developments have highlighted the importance of incorporating environmental influence, individual driving behavior, and attention mechanisms into predictive models. The use of graph attention networks, transformers, and large language models has also shown promising results in dynamic trajectory prediction and safety trajectory planning. Furthermore, the development of novel frameworks and architectures, such as those utilizing dual-phase and physics-informed approaches, has enhanced the robustness and performance of autonomous driving systems. Notably, papers such as 'Attention-Aware Multi-View Pedestrian Tracking' and 'GAMDTP: Dynamic Trajectory Prediction with Graph Attention Mamba Network' have demonstrated state-of-the-art performance in their respective areas, with the former achieving an IDF1 score of 96.1% on the Wildtrack dataset and the latter achieving superior accuracy in dynamic trajectory prediction on the Argoverse dataset.
Advances in Autonomous Driving and Traffic Prediction
Sources
Attention-Aware Multi-View Pedestrian Tracking
Towards An Efficient and Effective En Route Travel Time Estimation Framework
Vehicle Acceleration Prediction Considering Environmental Influence and Individual Driving Behavior
Data Scaling Laws for End-to-End Autonomous Driving
Planning Safety Trajectories with Dual-Phase, Physics-Informed, and Transportation Knowledge-Driven Large Language Models
GAMDTP: Dynamic Trajectory Prediction with Graph Attention Mamba Network
MIAT: Maneuver-Intention-Aware Transformer for Spatio-Temporal Trajectory Prediction
DyTTP: Trajectory Prediction with Normalization-Free Transformers
EP-Diffuser: An Efficient Diffusion Model for Traffic Scene Generation and Prediction via Polynomial Representations
PRIMEDrive-CoT: A Precognitive Chain-of-Thought Framework for Uncertainty-Aware Object Interaction in Driving Scene Scenario
Temporal-contextual Event Learning for Pedestrian Crossing Intent Prediction
CAFE-AD: Cross-Scenario Adaptive Feature Enhancement for Trajectory Planning in Autonomous Driving
Benchmarking Convolutional Neural Network and Graph Neural Network based Surrogate Models on a Real-World Car External Aerodynamics Dataset