To tackle the trajectory tracking problem and achieve high control accuracy in many actual nonlinear systems with unknown disturbance, a novel discrete-time extended state observer-(DESO) based model-free adaptive sliding mode control strategy with prescribed tracking performance is studied, which only relies on the input/output data of the system rather than explicit model information. Firstly, a compact-form dynamic linearization method is used to reconstruct the discrete-time nonlinear process, where the time-varying parameter linearly connected with the control input is obtained by an adaptive method and the unknown nonlinear term is estimated by a DESO. Then, by considering the prescribed performance and using an unconstrained vector transformed from the constrained tracking error, one model-free sliding mode controller is designed. In addition, a rigorous stability analysis is presented to show the boundedness of the sliding mode function and the prescribed transient-state and steady-state performance of the output tracking error. Finally, the simulations with comparing results verify the effectiveness and superiority of the developed control scheme. K E Y W O R D Sdiscrete-time extended state observer, discrete-time system, model-free adaptive control, prescribed performance, sliding mode control INTRODUCTIONWith the continuous development of engineering technologies, the complexity of the practical system is intensively increasing, which leads to accurate system modeling becoming one of the most difficult tasks. 1 Thus, the control strategies based on mathematical models are not suitable for such complicated systems. To overcome this problem, data-driven control is proposed where only the input and output data are used. Data-driven methods have been applied in many practical scenes, for instance, quadrotor vehicles, 2 automated vehicles, 3 continuum robots, 4 and other industrial process systems. 5,6 Among these data-driven methods, model-free adaptive control (MFAC) approach has aroused a lot of attention because only the input and output data are used without employing explicit or implicit knowledge of the mathematic

No abstract

To investigate the relationship between the Erk1/2 signal pathway and neuronal apoptosis in ischemic stroke rats. Male SD(Sprague Dawley) rats (n = 24) were randomly divided into three groups, each containing 8 rats: sham-operated group, MCAO(Midle cerebral artery oclusion) group, and MCAO + U0126 intervention group (U0126 group). In in vitro trial, primary cortical nerve cells were divided into three groups: control group, OGD(Oxygen and glucose deprivation) group, and U0126 intervention group (U0126 group). In vivo protein expression levels of Erk1/2, p-Erk1/2 and Bcl-2 were determined using western blot. The expressions of Bcl-2, Bcl-xl and Bax were assayed using immunohistochemical staining. Nerve cell mortality in cerebral tissue was detected using TUNEL staining. In in vitro trials, cell apoptosis was assayed with flow cytometry and LDH release. The activity of caspase-3 was determined. Nerve cell apoptosis was determined using Hoechst33258 staining method. In in vivo trial, it was found that the protein expression level of p-ERK1/2 in cerebral tissue in the MCAO group was significantly increased, when compared with that of the sham-operated group, while the protein expression level of p-Erk1/2 in the U0126 group was significantly lower than that in the MCAO group. The expression levels of Bcl-2 and Bcl-xl in the MCAO group were significantly lower than the corresponding expression levels in the sham-operated group, while the expressions of Bcl-2 and Bcl-xl in the U0126 group were significantly lower than those in MCAO group.In MCAO group, the expression of Bax was significantly higher than that in the sham-operated group, while Bax expression was higher in U0126 than in MCAO group. There were significantly higher number of dead nerve cells in MCAO group than in the sham-operated group, while nerve cell mortality in U0126 group was significantly lower than in MCAO group. In in vitro trials, flow cytometry revealed significantly higher apoptosis of OGD-treated nerve cells, relative to the control group. Nerve cells exposed to U0126 and treated with ODR (Oxygen-dependent repressor) were significantly decreased in population, when compared with single OGD treatment group. The LDH release level of nerve cells treated OGD was significantly increased, when compared with that of the control group.However, LDH release level of nerve cells treated with OGD after U0126 intervention was significantly decreased, relative to the single OGD treatment group.The dilution of nerve cell nucleus after OGD treatment was significantly increased, when compared with that of the control group. For nerve cells treated with ODR after U0126 intervention, the nuclear dilution was significantly decreased, relative to that of nerve cell nucleus in the single OGD treatment group. The OGD treatment led to significant increase in nerve cell caspase-3 activity, relative the control group. However, the caspase-3 activity of nerve cells treated with ODR after U0126 intervention was significantly decreased, when compared with single OGD treatm...

No abstract

No abstract

This is a repository copy of 2-D DOA estimation of incoherently distributed sources considering gain-phase perturbations in massive MIMO systems.

No abstract

Robust adversarial reinforcement learning is an effective method to train agents to manage uncertain disturbance and modeling errors in real environments. However, for systems that are sensitive to disturbances or those that are difficult to stabilize, it is easier to learn a powerful adversary than establish a stable control policy. An improper strong adversary can destabilize the system, introduce biases in the sampling process, make the learning process unstable, and even reduce the robustness of the policy. In this study, we consider the problem of ensuring system stability during training in the adversarial reinforcement learning architecture. The dissipative principle of robust H-infinity control is extended to the Markov Decision Process, and robust stability constraints are obtained based on L2 gain performance in the reinforcement learning system. Thus, we propose a dissipation-inequation-constraint-based adversarial reinforcement learning architecture. This architecture ensures the stability of the system during training by imposing constraints on the normal and adversarial agents. Theoretically, this architecture can be applied to a large family of deep reinforcement learning algorithms. Results of experiments in MuJoCo and GymFc environments show that our architecture effectively improves the robustness of the controller against environmental changes and adapts to more powerful adversaries. Results of the flight experiments on a real quadcopter indicate that our method can directly deploy the policy trained in the simulation environment to the real environment, and our controller outperforms the PID controller based on hardware-in-the-loop. Both our theoretical and empirical results provide new and critical outlooks on the adversarial reinforcement learning architecture from a rigorous robust control perspective.