Integrated sensor fault estimation and control for continuous-time switched systems: A new separation principle

Area of Science:

• Control Systems Engineering
• Fault Diagnosis and Tolerant Control
• Nonlinear Systems Analysis

Background:

• Switched systems are prevalent in many applications but are susceptible to sensor faults.
• Faults can degrade system performance and compromise stability.
• Existing methods often struggle with the complexity of fault estimation and control in switched systems.

Purpose of the Study:

• To develop a robust observer-based control strategy for switched systems experiencing sensor faults.
• To accurately estimate sensor faults and their impact on system states.
• To design a controller that compensates for these faults and ensures system stabilization.

Main Methods:

• Utilizing the mode-dependent average dwell time (MDADT) concept for switched system analysis.
• Applying Lyapunov stability theory to guarantee system stability.
• Developing a novel separation principle to formulate observer-based controller design using linear matrix inequalities (LMIs).

Main Results:

• A new separation principle is established, simplifying controller design from bilinear to linear matrix inequalities (LMIs).
• The proposed method effectively estimates sensor faults and compensates for their effects on state estimation.
• Demonstrated practicability and efficiency through simulations on a Highly Manoeuvrable Aircraft Technology (HiMAT) model, a DC-DC boost converter, and a general numerical example.

Conclusions:

• The developed observer-based control strategy provides an effective solution for state/fault estimation in switched systems with sensor faults.
• The use of MDADT and Lyapunov theory, combined with the new separation principle, offers a systematic approach to controller design.
• The method's efficiency and applicability are validated across diverse engineering examples.