Closed-Loop Stability of a Lyapunov-Based Switching Attitude Controller for Energy-Efficient Torque-Input-Selection During Flight

TL;DR

This paper introduces a Lyapunov-based switching attitude controller for UAVs that enhances energy efficiency and stability, validated through experiments showing significant control effort reduction during high-speed maneuvers.

Contribution

The paper develops a novel compound Lyapunov function for analyzing stability and region of attraction in a switching attitude control system for UAVs.

Findings

Reduces control effort by about 53% during yaw maneuvers

Provides a larger estimate of the system's region of attraction

Demonstrates stability and effectiveness through experimental flight data

Abstract

We present a new Lyapunov-based switching attitude controller for energy-efficient real-time selection of the torque inputted to an uncrewed aerial vehicle (UAV) during flight. The proposed method, using quaternions to describe the attitude of the controlled UAV, interchanges the stability properties of the two fixed points-one locally asymptotically stable and another unstable-of the resulting closed-loop (CL) switching dynamics of the system. In this approach, the switching events are triggered by the value of a compound energy-based function. To analyze and ensure the stability of the CL switching dynamics, we use classical nonlinear Lyapunov techniques, in combination with switching-systems theory. For this purpose, we introduce a new compound Lyapunov function (LF) that not only enables us to derive the conditions for CL asymptotic and exponential stability, but also provides us with an estimate of the CL system's region of attraction. This new estimate is considerably larger than those previously reported for systems of the type considered in this paper. To test and demonstrate the functionality, suitability, and performance of the proposed method, we present and discuss experimental data obtained using a 31-g quadrotor during the execution of high-speed yaw-tracking maneuvers. Also, we provide empirical evidence indicating that all the initial conditions chosen for these maneuvers, as estimated, lie inside the system's region of attraction. Last, experimental data obtained through these flight tests show that the proposed switching controller reduces the control effort by about 53%, on average, with respect to that corresponding to a commonly used benchmark control scheme, when executing a particular type of high-speed yaw-tracking maneuvers.