International Journal of Adaptive Control and Signal Processing最新文献

The accuracy of moving horizon state estimation significantly deteriorates when measurements are contaminated by outliers. Existing moving horizon estimation (MHE) methods that address this issue are restricted to linear systems. This paper proposes an outlier-robust MHE method to address the state estimation problem for nonlinear systems subject to measurement outliers. Specifically, at each sampling instant, the method solves a set of least-squares cost functions to eliminate potentially contaminated measurements one by one. The state estimate corresponding to the optimal cost is retained, and the process repeats as new information becomes available. Subsequently, the concept of uniform observability is introduced within this estimation framework. Applying the uniform observability property and choosing appropriate design parameters, the stability of the proposed estimator is proved. Additionally, a robustness condition is derived, ensuring that the estimator remains resilient to outliers, provided they are sufficiently large. Finally, the parameter design method to achieve stability and robustness in the estimator implementation is presented. The simulation results show the effectiveness of the proposed estimation approach in case the measurements are contaminated.

A Double Internal Loop Recurrent Neural Network (DILRNN) with an adaptive learning rate is proposed for the modeling and control of non-linear dynamical plants. The structure of DILRNN is an alteration of the Fully Connected Recurrent Neural Network (FCRNN). DILRNN contains three feedback loops taken primarily from the context layer to the hidden layer, the time delay of the output layer to the hidden layer, and the time delay of output to output. The parameters of DILRNN are updated using the gradient descent-based dynamic Back-Propagation (BP) algorithm. The Adaptive Learning Rate (ALR) scheme is implemented to ensure that the learning rate value is determined properly in each iteration and improves the performance of the learning algorithm. The effectiveness of the suggested strategy is assessed by considering both the Multiple Input Single Output (MISO) and Multiple Input Multiple Output (MIMO) systems and comparing them with state-of-the-art methods. The simulation outcomes show that the proposed model outperforms the Feed Forward Neural Network (FFNN) and other Recurrent Neural Network (RNN) models, which were taken into evaluation in terms of their output and error. Next, a DILRNN controller is implemented using the proposed model to control non-linear dynamical systems. The controller's response is evaluated both in the absence and presence of a disturbance signal to assess the recovery capability of the controller. The response and error of the proposed controller are compared with other neural network controllers and elementary PID controllers.

This paper investigates the adaptive neural network- controlled course tracking of an unmanned surface vehicle (USV) with quantization of signal input. As a first step, the characteristics of the ship's rudder servo system are fully considered and combined with the mathematical representation of the ship's heading control system. This is done to develop a nonlinear third-order response model. The Radial Basis Function (RBF) neural network is constructed to estimate and approximate the unknown functions within a mathematical model of the system, and nonlinear damping terms are employed to counteract external disturbances. Subsequently, a design method for a neural network adaptive quantization controller is proposed. This controller can enable real-time learning and adjustment to address performance degradation caused by signal quantization errors. Based on the Lyapunov theorem, the designed controller has been validated for its dynamic response capability and system stability, ensuring long-term reliable and stable operation. In addition, semiglobally, uniformly bound signals are used in closed-loop systems. Tracking errors are lowered through parameter tuning to trim levels arbitrarily. As a final result, simulation results confirmed the effectiveness and feasibility of the RBF neural network-based adaptive quantification control method for USVs.

This paper mainly investigates the joint estimation of the parameters and states for the fractional-order Wiener state space model. Based on the Kalman filter principle, a generalized recursive least squares algorithm with a forgetting factor is proposed. In addition, the filtering-based generalized recursive least squares algorithm is presented, which reduces the influence of colored noise on the parameter estimation. A gradient identification algorithm is introduced to estimate the order of the fractional-order. Under the persistent excitation conditions, the analysis indicates that the proposed parameter estimation algorithm can estimate the fractional-order Wiener state space system. A simulation example is given to confirm that the proposed algorithms are effective.

Task allocation in multi-robot systems has been a critical area of research, with applications spanning various industries, such as logistics, agriculture, and manufacturing. The allocation of tasks to multi-robots improves the system performance, which generally minimizes total resource consumption or cost needed for performing a group of tasks. In dynamic multi-robot systems, efficient task allocation is critical for optimizing system performance, especially in response to environmental changes like faults or the actions of other robots. Therefore, a new approach called reptile honey badger optimization algorithm_deep quantum neural network (RHBA_DQNN) is framed for task allocation in multi-robot systems. At first, the tasks are grouped utilizing the fuzzy local information C-means (FLICM) clustering model. Then, the assignment of tasks for the group of robots is conducted using the devised RHBA, where monetary cost, distance, time, and completion time are considered objective functions. The proposed RHBA is the combination of the reptile search algorithm (RSA) and honey badger algorithm (HBA). Finally, the penalty cost is decided based on the deep quantum neural network (DQNN). Moreover, the RHBA_DQNN has obtained a minimum overall cost, execution time, distance, and monetary cost of 81.251, 9.99, 1.600, and 0.249, respectively.

This article presents an integral reinforcement learning-based optimal formation tracking scheme for multiple quadrotor unmanned aerial vehicles (QUAVs) experiencing nonlinear coupled dynamics and subject to constraints. We use multilayer neural networks (MNN) within an actor-critic framework where the MNN weights are tuned using singular value decomposition (SVD) of the activation function gradient to approximate optimal control policy via backstepping. Additionally, barrier Lyapunov functions (BLF) are introduced to ensure set invariance, thereby maintaining the quadrotors within a defined safety space due to constraints. A novel weight update law for each layer is derived using the HJB approximation error and control input error. Stabilizing terms for the output layer, obtained through Lyapunov analysis, are included to enhance stability and ensure the boundedness of the system. To improve performance on multitasking missions and address the issue of catastrophic forgetting, online continual learning is incorporated in each layer of actor-critic MNNs. Moreover, this method is applied for leader–follower formation using spherical coordinates. The control objectives for the followers involve tracking the leader with the desired separation, angle of incidence, and bearing through auxiliary velocity control. The simulation results indicate potential improvements over traditional methods.

This article presents a novel Adaptive Double Integral-Type Second-Order Terminal Sliding Mode (ADITSOTSM) control policy for favorable and swift stabilization of the disturbed uncertain Cyber-Physical Systems (CPSs) in finite time. The offered double integral-type switching surface elides the approaching step and meliorates the system's robustness. Real-time adaptive laws are expanded for coping with unwanted disturbances, parametric uncertainties, and cyber attacks on sensors and actuators, so that their upper bounds designation is not indispensable. The recommended strategy guarantees the disturbed uncertain CPSs robust functionality against rigorous cyber attacks on sensors and actuators. In addition, it provides swift response, smooth and robust control, high carefulness and more flexible operation, finite time convergence, without chattering, and no transient fluctuations. Comparing the simulation results with conventional, adaptive integral-type sliding mode, and state-feedback control depicts the introduced technique's high success and effectiveness.

In this paper, we present a new scheme to design an adaptive neural backstepping tracking controller for a class of stochastic nonlinear systems with unknown input constraints including saturation and dead-zone. The control design is achieved by using certain auxiliary techniques. More specifically, some piecewise injective differential functions are constructed to approximate these nonlinearity constraints with a bounded approximation error. A new dimension reduction inequality related the norm of basis vectors of radial basis functions is established to design the virtual signals. Some negative power exponential functions and the inverse functions of the above injective differentiable functions are introduced to design the adaptive laws and controller, respectively. These techniques can enlarge the nonlinear systems currently studied by backstepping approach and neural networks. Furthermore, it is shown that all the signals in the closed-loop system are semi-globally uniformly, ultimately bounded in probability and the tracking error converges to zero. Meanwhile, the effectiveness of the proposed controller is demonstrated in the simulation study.

This article investigates the problem of event-triggered output feedback for a class of uncertain nonlinear time-delay systems. Essentially different from previous works, the system nonlinearities satisfy certain bounded condition depending on an unknown growth coefficient, the input–output function, the unmeasurable states, and an unknown time-varying delay. The complicated nonlinearities pose a substantial challenge in solving this problem. Employing the dynamic gain technique, we construct a new adaptive event-triggered controller via output feedback, where the event-triggering mechanism is a combination of event-triggered and time-triggered. We prove that all states of the concerned system globally converge to zero, and Zeno phenomenon is excluded. We further extend the results to nonlinear time-delay systems with input matching uncertainty. Numerical examples are given to verify the validity of the theoretical results.