Longitudinal Control
Longitudinal control, focusing on speed and distance management in vehicles, is a critical area of research in autonomous driving and related multi-agent systems. Current research emphasizes developing robust and safe longitudinal control algorithms, employing techniques like model predictive control (MPC), deep reinforcement learning (DRL), and PID controllers, often enhanced with control barrier functions for safety guarantees. These advancements aim to improve vehicle safety, efficiency, and coordination, particularly in challenging scenarios such as high-density traffic and emergency braking, with applications ranging from autonomous vehicles to automated material handling systems.
Papers
Deep Reinforcement Learning for Advanced Longitudinal Control and Collision Avoidance in High-Risk Driving Scenarios
Dianwei Chen, Yaobang Gong, Xianfeng Yang
PlanNetX: Learning an Efficient Neural Network Planner from MPC for Longitudinal Control
Jasper Hoffmann, Diego Fernandez, Julien Brosseit, Julian Bernhard, Klemens Esterle, Moritz Werling, Michael Karg, Joschka Boedecker