Reducing real-time complexity via sub-control Lyapunov functions: from theory to experiments

TL;DR

This paper introduces Sub-control Lyapunov functions (SCLFs) as a computationally efficient alternative to traditional control Lyapunov functions, reducing online complexity in control systems through linear programming and validated by drone experiments.

Contribution

It proposes a novel SCLF framework that simplifies online control computation by using linear programming with basis functions, bridging theory and practical experiments.

Findings

SCLFs can be computed via linear programming with infinite constraints.

The approach significantly reduces online computational complexity.

Experiments on drone control validate the effectiveness of the method.

Abstract

The techniques to design control Lyapunov functions (CLF), along with a proper stabilizing feedback, possibly in the presence of constraints, often provide control laws that are too complex for proper implementation online, especially when an optimization problem is involved. In this work, we show how to acquire an alternative, computationally attractive feedback. Given a nominal CLF and a nominal state feedback, we say that a different positive definite function is a Sub-control Lyapunov function (SCLF) if its Lyapunov derivative is negative-definite and bounded above by the Lyapunov derivative of the nominal function with the nominal control. It turns out that if we consider a family of basis functions, then a SCLF can be computed by linear programming, with an infinite number of constraints. The idea is that although the offline computational burden to achieve the new controller and solve the linear program is considerable, the online computational burden is drastically reduced. Comprehensive simulations and experiments on drone control are conducted to demonstrate the effectiveness of the study.