This study explores a nonsingular predefined-time control approach using dynamic surface control (DSC) for nonlinear multi-agent networks. First, under the predefined-time stability criterion, a smooth controller is realised, which allows for the system convergence time to be set beforehand. Furthermore, the nonlinear decomposition technique is used to address input quantisation, and the coupling problem between multi-agent systems and nonlinear dynamics is solved with RBF neural network approximator. Moreover, prescribed-time performance function is devised for tracking error performance convergence, while eliminating the limitations of the initial error condition. Furthermore, the difficult problem of integrating dynamic surface control into predefined time stability is addressed, dealing with the problem of explosion of complexity. It is finally demonstrated that the suggested controller can ensure the closed-loop system's predetermined time stability based on Lyapunov stability analysis, and the final simulation confirms the efficacy of the suggested approach.
In this paper, the synchronisation problem for a category of complex dynamical networks (CDNs) incorporating dynamic relay links (DRLs) with input-output function is investigated. In contrast to the current literature, this paper introduces the concept of a relay-conversion relationship to describe the connections between nodes and proposes a complex network model composed of a DRL subsystem and a node subsystem. In this model, these two subsystems are coupled by their respective outputs, which demonstrates the unique input-output characteristic and function of the DRLs. Subsequently, to realise node synchronisation, the node controller and the DRL coupling term are designed based on Lyapunov stability theory. Furthermore, under the coupling effect of the node subsystem, the DRL synchronisation is also achieved while the nodes achieve synchronisation. Finally, by using a numerical simulation example, the viability of the control scheme proposed in this paper is confirmed.
This paper mainly studies the problem of fixed-time state-constrained vector L2-gain and fixed time stability for switched nonlinear systems. First, the definition of fixed-time state-constrained vector L2-gain is proposed. This definition requires each active subsystem to have a fixed-time vector L2-gain during its active time intervals. And the energy is allowed to rise at each switching time. Second, sufficient conditions for the fixed-time state-constrained vector L2-gain are provided, and based on this condition, a state-dependent switching law is designed to achieve fixed-time state-constrained vector L2-gain. Finally, even if none of the active subsystems has a fixed-time state-constrained vector L2-gain, the fixed time stability of the closed-loop system is obtained. The effectiveness of the proposed method is verified through two examples.
This paper presents a contact-aware adaptive finite-time sliding mode control framework for a miniature wall-climbing aerial robot (WCAR) designed for autonomous in-contact inspection in structural health monitoring (SHM). The WCAR is a lightweight quadrotor with passive wheels, enabling smooth transitions between free flight, wall contact, and vertical climbing. Unlike conventional methods that assume known interaction forces or only guarantee asymptotic convergence, the proposed approach combines integral terminal sliding mode control with a finite-time disturbance observer and online friction adaptation to handle inertia uncertainty and unknown wall friction. Phase-specific sliding surfaces ensure finite-time convergence across operational modes. Closed-loop stability is established via Lyapunov analysis. Simulation and real-time experiments on a custom 61-g platform validate robustness against friction variations, inertia mismatch, and disturbances, demonstrating improved convergence and performance over existing adaptive terminal sliding mode control methods.
This study proposes a novel discrete-time implementation to eliminate numerical chattering in the digital twisting controller. Bypassing the computational complexity of state-of-the-art implicit solvers, the methodology employs an efficient hybrid strategy. First, rigorous discrete-time Lyapunov analysis proves that a semi-implicit Euler discretization achieves finite-time convergence toward an analytically defined spatial region representing the exact bounds of numerical chattering. Upon entering this boundary, the control logic switches to a purely linear deadbeat formulation, driving the states exactly to the origin in a maximum of two steps without chattering. Robustness and ultimate tracking bounds are analytically established under parametric uncertainties, matched disturbances, and measurement noise. Simulations validate that this hybrid approach matches the high-precision convergence of advanced implicit methods while offering a simpler implementation.
This paper presents a data-driven risk-aware model predictive control (MPC) framework for discrete-time linear systems under process noise. To improve safety and performance predictability, conditional value-at-risk (CVaR) is utilised in the cost function and constraint of the MPC problem. Leveraging CVaR instead of the expected value avoids fluctuation of performance and safety metrics under noise. First, a computationally efficient solution is provided for the multistage CVaR optimisation problem. This is achieved using the dual representation of multi-stage risk measures and data-driven ambiguity sets from which the multi-stage CVaR problem is cast as a tractable semidefinite programming (SDP) optimisation. Then, the CVaR-based MPC problem is cast as an SDP optimisation problem. The recursive feasibility and risk-aware exponential stability of the presented risk-aware MPC are demonstrated through rigorous theoretical analysis by considering the disturbance feedback policy parameterisation. Finally, two numerical examples are given to elucidate the efficacy of the presented method.
