Unmanned Vehicle Navigation Based on Speed-Adjusted Control Barrier Functions within a Model Predictive Control Framework

In this paper, we propose a Speed-Adjusted Control Barrier Function based on Model Predictive Control (MPC-SDCBF) to address the conflict between safety guarantees and performance objectives in autonomous vehicle navigation. Specif-ically, we introduce a SDCBF for discrete-time systems, which dynamically links the robot's real-time velocity to its safety set. This approach transforms the safety set into a dynamic set that varies adaptively with the robot's velocity. By employing this method, the range of the safety set can be flexibly adjusted to balance the aforementioned conflict. Subsequently, we integrate SDCBF constraints into an MPC framework to solve the single-vehicle navigation problem while avoiding collisions with both static and dynamic obstacles. Finally, the effectiveness of the proposed method is validated through simulation experiments.