Data-Driven Dead-Zone Compensation via Projection in Predictive Control Setting

Fixed-Time Stable Fault-Tolerant Control of Underactuated Hovercraft via Physics-Informed Neural Adaptation

This study addresses the trajectory tracking control problem for an underactuated hovercraft subject to additive bias and multiplicative loss-of-effectiveness thruster faults under environmental disturbances. In these systems, actuator degradation structurally breaks the differential flatness mapping, driving nominal controllers to generate control actions that induce severe actuator saturation and cause instability. To resolve this challenge, a hierarchical physics-informed neural adaptive control (PINAC) framework is proposed. First, a gated-recurrent-unit physics-informed neural observer (PINO) is designed to isolate thruster faults from exogenous hydrodynamic disturbances. Second, a constrained Safe-TD3 reinforcement learning agent functions as a supervisor, computing an online dilation factor to slow down the mission timeline, thereby reconfiguring the reference trajectory to accommodate degraded actuator boundaries. Third, a low-level non-singular terminal sliding mode (NTSM) controller is implemented as a tracking-guarantee layer. Unlike classical asymptotic schemes where convergence is only achieved as time approaches infinity, or finite-time controllers where the settling time depends on the initial state, the proposed PINAC framework guarantees practical fixed-time stability, ensuring that the settling-time bound is independent of initial conditions. Simulation results demonstrate that the designed controller prevents actuator saturation, provides smooth trajectory adjustment, and reduces tracking errors under severe composite faults.

Data-Driven Disturbance-Observer-Based Actuator-Space Control of a Dual-Axis Thrust-Vectoring Platform

Electromechanical thrust-vector control systems are subject to friction, backlash, and configuration-dependent dynamics that are difficult to model explicitly, motivating controllers that adapt from operational data without requiring an identified plant. This paper proposes a data-driven, disturbance-observer-based controller in actuator space for a dual-axis thrust-vectoring platform. The controller operates on a sliding surface defined over the actuator tracking error and uses an adaptive input matrix gain together with a lumped disturbance observer, both of which are updated from encoder and control data through gradient descent applied to a joint identification loss. The Newton–Euler equations of the mechanical system are projected onto the actuator coordinates through the angular-velocity map, yielding a structurally well-posed actuator-space model, which is used to design the adaptive gain; nonetheless, the dynamic model is never evaluated online. The resulting controller requires only encoder measurements of actuator displacements, a reference trajectory computed from the platform geometry, and bounded normalized commands; the mechanism Jacobian and the inertia and Coriolis matrices are not evaluated online. Boundedness of the adaptive gain and disturbance estimate is established, and uniform ultimate boundedness of the tracking error follows under a mild alignment condition. Simulation results on a coupled nonlinear plant and hardware-in-the-loop experiments on a dual-channel actuator testbed are presented, comparing the proposed controller against PID, super-twisting, unit-vector sliding mode, and MFAC baselines on a circular thrust-vector reference.

Admissibility-Preserving Control for Strict-Feedback Nonlinear Systems with Asymmetric Actuator Constraints

This paper develops Admissibility-Preserving Control (APC), a realization-centered safety-critical control framework for strict-feedback systems subject to asymmetric actuator limits, time-varying output constraints, and actuator-rate limitations. APC denotes the overall control architecture, whereas an Admissibility-Preserving Input Realization (APIR) denotes its constraint-realization module. Therein, the APIR dynamically generates the physical plant input while rendering its prescribed asymmetric actuator set forward invariant. In contrast to algebraic clipping and post-design saturation compensation, the actuator limits are embedded directly in a continuously differentiable dynamic realization with user-selectable regularity and interpretable tuning parameters. The APIR is integrated with recursive backstepping by treating the realized plant input as an additional state. The resulting design does not require an input-to-state stability assumption on the uncontrolled plant. Instead, the nonlinear drift terms are compensated recursively, subject to an explicit compatibility condition between the desired motion, the available control authority, and the APIR interior gain. The framework is further extended to time-varying output-safe tracking through a smooth asymmetric logarithmic barrier coordinate and its associated Lyapunov function and to simultaneous actuator-magnitude and rate constraints through a cascaded APIR. Rigorous Lyapunov and invariance analyses establish regional asymptotic tracking, forward invariance of the compatible admissible sets, and boundedness of all closed-loop signals. Numerical studies illustrate asymmetric actuator utilization, output-safety preservation, and magnitude-rate constraint enforcement.

Rate-Dependent Hysteresis Modeling and Hybrid Inverse Compensation Control for Piezoelectric Actuators

Highlights What are the main findings? A seventh-order, frequency-dependent rising/falling branch model is selected using validation error and endpoint-stability criteria. A direct inverse feedforward path is integrated with disturbance-observer-based adaptive sliding-mode feedback and verified on a laboratory piezoelectric stack platform. What are the implications of the main findings? The direct inverse model strongly reduces hysteresis nonlinearity, while the feedback loop compensates residual model errors, bounded disturbances, and unmodeled dynamics. The hybrid architecture improves tracking beyond feedforward-only compensation while smooth tanh switching and projection adaptation suppress chattering and parameter drift. Abstract Piezoelectric ceramic actuators are widely used in precision positioning and sensor-integrated micro-motion systems, but their accuracy is limited by asymmetric, rate-dependent hysteresis and by residual disturbances that remain after feedforward linearization. This study develops a self-contained modeling and control framework that combines an explicit rising/falling branch polynomial model, frequency-dependent coefficient maps, direct inverse feedforward compensation, and disturbance-observer-based adaptive sliding-mode feedback. The actuator is represented as a multilayer piezoelectric stack coupled to an equivalent electrical-mechanical-sensing plant. A branch-state logic resolves the multivalued inverse mapping, and a numerical order-sensitivity study shows that the seventh-order model provides the lowest validation RMSE while avoiding the endpoint growth observed at higher orders. Laboratory measurements at 1, 5, 10, 20, 50, and 100 Hz, together with attenuated, triangular, random-amplitude, step, and 2 Hz sinusoidal tests, are used for validation. The proposed branch model reduces static maximum relative fitting error from 4.50–6.21% for the classical P-I model to 1.28–2.58%. Direct inverse compensation reduces linearity error from 8.56–13.88% to 0.53–1.024%, and the hybrid controller achieves a 1% settling time of 8.6 ms, a maximum tracking error of 0.0051 micrometers, and an RMSE of 0.0012 micrometers. The results demonstrate an embedded-oriented compromise between model accuracy, online computational simplicity, and robust closed-loop precision.

Neuromorphic Automatic Reset PI Controller for Simultaneous Anti-Windup and Dead-Zone Compensation: Initial Results

Actuator saturation in systems regulated by proportional-integral (PI) controllers causes integrator ‘windup’ and, consequently, a deterioration in performance. Conventional PI control configurations with automatic reset address this problem by introducing a saturation model that limits the integral action. This article proposes an anti-windup PI control scheme based on neuromorphic control (NC), in which conventional saturation is replaced by a neuron. This element fulfils a dual function: thanks to its pulse-based operation, it inherently imposes saturation when the input exceeds the pulse amplitude, whilst mitigating the effects of the dead zone without requiring explicit modelling. By addressing both non-linearities through a single element, the overall control architecture remains remarkably simple. Simulations results on a direct current (DC) motor demonstrate the simplicity and effectiveness of the proposed strategy in a non-linear actuator with dead-zone and saturation.

Data-driven stochastic model predictive control for spacecraft attitude tracking