On the Equivalence of Time-Varying CBF-Based Control and Prescribed Performance Control : Conversion and Qualitative Comparison

Sammanfattning: These days, a wide range of autonomous systems, such as automobiles, delivery drones, and embedded household systems, are becoming more and more common in our society. This trend is projected to continue in the future. To effectively manage these dynamic systems, ensuring their safe operation is crucial for the well-being of our lives. Control of safety-critical systems has gained significant attention in recent years, particularly in the field of nonlinear control. While the mathematical tools for characterizing safety are well-established, there are still numerous challenges to be addressed when it comes to developing methodologies for synthesizing nonlinear control systems. This report investigates the similarity between the two control schemes, the prescribed performance control and control barrier function. Its purpose is to shed light on the development of control methodology in safetycritical systems. While both methods have been successfully constructed and developed recently, there is no existing report that clarifies their similarities. To gain a deeper understanding of the latest safety-critical control and investigate these similarities, this report aims to provide interesting insights and contribute to the further development of methodology. The key insight arises from the fact that the prescribed performance control can be considered a method based on barrier functions. Consequently, it can be regarded as a control barrier-based controller. In order to demonstrate the similarities and make a comparison between the two, a unified problem setting is presented. Once we have properly converted the problem, we can proceed with a comparison using numerical simulations. The results presented in this report demonstrate that the prescribed performance controller can be implemented using separate reciprocal CBF methods. Furthermore, it shows that the performance achieved is comparable to that of the CLF-CBF QP, which utilizes optimization techniques to ensure stability and safety requirements. These findings raise new questions regarding the relationship between these two approaches. Ultimately, the report delves into a deeper understanding of how model-free methods achieve superior performance compared to model-based methods that heavily rely on optimization.