High-Speed Highway Scene Prediction Based on Driver Models Learned From Demonstrations

David Sierra González (1) , Jilles Steeve Dibangoye (1, 2, 3) , Christian Laugier (1)

1 CHROMA - Robots coopératifs et adaptés à la présence humaine en environnements dynamiques
2 INSA Lyon - Institut National des Sciences Appliquées de Lyon
3 CITI - CITI Centre of Innovation in Telecommunications and Integration of services

One of the key factors to ensure the safe operation of autonomous and semi-autonomous vehicles in dynamic environments is the ability to accurately predict the motion of the dynamic obstacles in the scene. In this work, we show how to use a realistic driver model learned from demonstrations via Inverse Reinforcement Learning to predict the long-term evolution of highway traffic scenes. We model each traffic participant as a Markov Decision Process in which the cost function is a linear combination of static and dynamic features. In particular, the static features capture the preferences of the driver while the dynamic features, which change over time depending on the actions of the other traffic participants, capture the driver's risk-aversive behavior. Using such a model for prediction enables us to explicitly consider the interactions between traffic participants while keeping the computational complexity quadratic in the number of vehicles in the scene. Preliminary experiments in simulated and real scenarios show the capability of our approach to produce reliable, human-like scene predictions.

Mots clés: Inverse Reinforcement Learning; Intelligent Vehicles; Motion Prediction; Motion Planning; Motion Planning in Dynamic Environments; Planning-Based Motion Prediction

Domaines: Intelligence artificielle [cs.AI]; Robotique [cs.RO]; Machine Learning [stat.ML]

David Sierra González, Jilles Steeve Dibangoye, Christian Laugier. High-Speed Highway Scene Prediction Based on Driver Models Learned From Demonstrations. Proceedings of the 2016 IEEE 19th International Conference on Intelligent Transportation Systems (ITSC 2016), Nov 2016, Rio de Janeiro, Brazil. ⟨10.1109/ITSC.2016.7795546⟩. ⟨hal-01396047⟩