Behavior Prediction
Behavior prediction aims to forecast the future actions of agents, such as humans, vehicles, or even horses, leveraging diverse data sources like sensor readings, video footage, and social media interactions. Current research emphasizes developing sophisticated models, including neural networks (e.g., recurrent and convolutional architectures, transformers), graph neural networks, and probabilistic methods (e.g., normalizing flows), to capture complex spatio-temporal dependencies and account for uncertainty in predictions. These advancements are crucial for improving safety and efficiency in various applications, including autonomous driving, human-robot interaction, and personalized healthcare, by enabling proactive and adaptive systems.
Papers
Toward Unified Practices in Trajectory Prediction Research on Drone Datasets
Theodor Westny, Björn Olofsson, Erik Frisk
RAG-based Explainable Prediction of Road Users Behaviors for Automated Driving using Knowledge Graphs and Large Language Models
Mohamed Manzour Hussien, Angie Nataly Melo, Augusto Luis Ballardini, Carlota Salinas Maldonado, Rubén Izquierdo, Miguel Ángel Sotelo