Advances in Autonomous Vehicle Control

The field of autonomous vehicle control is moving towards more sophisticated and efficient control strategies. Researchers are exploring novel approaches that combine traditional control methods with machine learning and optimization techniques to improve the performance and robustness of autonomous vehicles. One notable trend is the integration of model predictive control (MPC) with other control methods, such as deep reinforcement learning and linear parameter varying (LPV) control, to achieve better tracking accuracy and adaptability. Another area of focus is the development of reactive controllers that can handle complex scenarios and avoid common issues such as dead-ends and incomplete boundaries. Noteworthy papers include DTR, which presents a reactive controller that combines Delaunay triangulation with track boundary segmentation to achieve 70% faster lap times, and LoL-NMPC, which proposes a novel NMPC formulation that incorporates low-level controller dynamics to minimize trajectory tracking errors.

Sources

DTR: Delaunay Triangulation-based Racing for Scaled Autonomous Racing

LoL-NMPC: Low-Level Dynamics Integration in Nonlinear Model Predictive Control for Unmanned Aerial Vehicles

Autonomous Vehicle Lateral Control Using Deep Reinforcement Learning with MPC-PID Demonstration

Real-Time LPV-Based Non-Linear Model Predictive Control for Robust Trajectory Tracking in Autonomous Vehicles

Tire Wear Aware Trajectory Tracking Control for Multi-axle Swerve-drive Autonomous Mobile Robots

Built with on top of