Hybrid Control of ADT Switched Linear Systems subject to Actuator Saturation

This paper develops a hybrid output-feedback control framework for average dwell-time (ADT) switched linear systems subject to actuator saturation. The considered subsystems may be exponentially unstable, and the saturation nonlinearity is explicitly handled through a deadzone-based representation. The proposed hybrid controller combines mode-dependent full-order dynamic output-feedback controllers with a supervisory reset mechanism that updates controller states at switching instants. By incorporating the reset rule directly into the synthesis conditions, switching boundary constraints and performance requirements are addressed in a unified convex formulation. Sufficient conditions are derived in terms of linear matrix inequalities (LMIs) to guarantee exponential stability under ADT switching and a prescribed weighted ${\cal L}_2$-gain disturbance attenuation level for energy-bounded disturbances. An explicit controller construction algorithm is provided based on feasible LMI solutions. Simulation results demonstrate the effectiveness and computational tractability of the proposed approach and highlight its advantages over existing output-feedback saturation control methods.

💡 Research Summary

The paper presents a novel hybrid output‑feedback control framework for continuous‑time switched linear systems that are subject to actuator saturation and governed by an average dwell‑time (ADT) switching rule. The authors first model the saturation nonlinearity using a dead‑zone representation, i.e., sat(u)=u−dz(u), where dz(u) is a bounded function that captures the difference between the actual input and its saturated value. This representation enables the saturation to be incorporated into a linear fractional transformation (LFT) of the plant, preserving convexity in the subsequent synthesis.

The core of the contribution lies in combining mode‑dependent full‑order dynamic output‑feedback controllers with a supervisory reset mechanism. At each switching instant the controller state is instantaneously reset by a matrix Δij that maps the pre‑switching controller state to a new initial condition for the post‑switching mode. By embedding this reset rule directly into the Lyapunov‑based stability analysis, the authors avoid the bilinear matrix inequalities (BMIs) that typically arise from switching‑induced boundary conditions.

Stability and performance are addressed simultaneously using a multiple Lyapunov function (MLF) approach. For each subsystem i a positive‑definite matrix Pi defines a quadratic Lyapunov function Vi(x)=xᵀPi x. The MLF conditions require (i) a uniform decay rate λ0, (ii) a bound μ such that Vi≤μ Vj for any pair of modes, and (iii) a weighted L2‑gain bound γ for energy‑bounded disturbances. The average dwell‑time τa must satisfy τa≥(ln μ)/λ0, which is directly linked to the MLF parameters. All these requirements are expressed as a single block‑matrix inequality that is linear in the decision variables (Pi, controller gains K_i, L_i, and reset matrices Δij). Consequently, the synthesis problem reduces to a set of linear matrix inequalities (LMIs) that can be solved efficiently with standard semidefinite programming tools.

The design procedure proceeds as follows: (1) formulate the closed‑loop augmented system that includes the plant, the dead‑zone model, and the dynamic controller; (2) construct the block LMI that encodes exponential stability, the switching‑boundary condition (through Δij), and the weighted L2‑gain; (3) solve the LMI feasibility problem; (4) extract the controller matrices and reset maps from the solution. The authors provide an explicit algorithm that maps feasible LMI solutions to a concrete hybrid controller implementation.

Simulation studies involve a fourth‑order switched system composed of two exponentially unstable subsystems. Saturation limits are set to ±1.0, and an external disturbance with bounded L2‑energy (‖w‖₂≤0.5) is injected. With an average dwell‑time of τa=0.5 s and a desired L2‑gain γ=1.2, the synthesized hybrid controller achieves exponential convergence of the state for all admissible switching sequences and guarantees that the output energy never exceeds γ² times the disturbance energy. Comparative simulations show that (i) a conventional output‑feedback design that ignores saturation leads to divergence, and (ii) an output‑feedback design without state resets violates the dwell‑time bound and exhibits degraded performance. These results demonstrate that the proposed method significantly reduces conservatism while maintaining computational tractability.

In conclusion, the paper delivers a systematic, convex LMI‑based synthesis method for hybrid output‑feedback control of ADT‑switched linear systems with actuator saturation. By jointly handling saturation nonlinearity, switching‑induced state jumps, and disturbance attenuation, the approach extends existing results that either neglect saturation, rely on full‑state feedback, or suffer from BMI‑type complexity. Future work suggested includes tighter dead‑zone approximations, time‑varying dwell‑time policies, and real‑time implementation of the reset mechanism.