Information-Theoretic Sensor-Based Predictive Control for Autonomous Vehicle Navigation: A Proof of Concept

David Perez-Morales (1, 2, 3) , Vincent Frémont (1, 2, 3)

1 ECN - École Centrale de Nantes
2 LS2N - Laboratoire des Sciences du Numérique de Nantes
3 LS2N - équipe ARMEN - Autonomie des Robots et Maîtrise des interactions avec l’ENvironnement

This paper explores the feasibility of an Information-Theoretic Sensor-Based Predictive Control (IT-SBPC) approach for autonomous navigation in presence of pedestrians. Our technique relies solely in sensor data expressed relative to the vehicle and therefore no localization is inherently required. By combining the advantages of the information-theoretic framework and sensor-based formalism, the proposed technique drives the vehicle safely and smoothly towards the desired goal. Several real-time simulated scenarios, showing that the car is able to reach the goal with centimeter-level accuracy, validate the effectiveness of our approach.

Mots clés: Location awareness; Navigation; Optimal control; Real-time systems; Trajectory; Safety; Automobiles

Domaines: Robotique [cs.RO]; Vision par ordinateur et reconnaissance de formes [cs.CV]; Intelligence artificielle [cs.AI]

David Perez-Morales, Vincent Frémont. Information-Theoretic Sensor-Based Predictive Control for Autonomous Vehicle Navigation: A Proof of Concept. 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), Sep 2021, Indianapolis, United States. pp.879-884, ⟨10.1109/ITSC48978.2021.9564855⟩. ⟨hal-03413632⟩