Answers are grounded in this paper and cite their sources.
This article proposes a safety-critical control framework for autonomous vehicles operating in dynamic environments. Dynamic obstacles are represented by online-updated oriented minimum bounding ellipses (MBEs), providing a compact and geometry-aware approximation of irregular obstacle shapes. Abarrier function is constructed from the minimum Euclidean distance between the vehicle and each fit ellipse, while the effect of the estimated obstacle motion is explicitly retained in the barrier derivative. The resulting dynamic control barrier function (DCBF) is incorporated into model predictive control (MPC) through discrete-time control barrier function (CBF) constraints. To balance safety conservatism and optimization feasibility under varying vehicle-obstacle interactions, a bounded sigmoid-based mechanism adjusts the CBF decay rate online. In addition, a constant-velocity Kalman filter (CV-KF) estimates obstacle positions and velocities from noisy measurements and propagates the estimated states over the MPC prediction horizon. The simulation and experimental results demonstrate that the proposed approach achieves safe and efficient obstacle avoidance in the tested dynamic scenarios.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: