Search

In this paper, we consider a constrained model predictive control design for an Elbow Joint Orthosis Robot. The challenge of the robotic device control is to achieve passive and resistive rehabilitation movements with hard and soft constraints. A fast quadratic programming solver based on the Hildreth method is developed to meet the short real-time implementation sampling time. The online optimization solver based on the Hildreth method is compared with two other solvers proving its superiority in terms of computation time in the simulation environment. Simulation results show successful set point tracking and constraints satisfaction. An embedded version of constrained model predictive control using the Hildreth quadratic programming solver is implemented on an Arduino Due microcontroller. The results are satisfactory, and motion tracking errors and inputs variance are minimized. Less effort, optimal performances, and absence of overshoot and control signal oscillations are considered to evaluate the different studied motions. A lower computational cost is obtained taking into account the trade-off between the control effort and the closed-loop performances.

This article is concerned with the design of wireless-networked control framework based on Internet of Things technology and predictive control strategy to remote control a thermal benchmark system. In order to improve the control performance of systems, an autonomous real-time solution is proposed for handling network problems. The adopted control strategy is divided into two cooperative parts under a master–slave architecture, in which two STM32 microcontrollers are investigated. The slave board is connected closely to the process and the master one is a distant controller. The microcontrollers communicate wirelessly through the Transmission Control Protocol/Internet Protocol. In the master board, a model predictive output-estimator-based controller is designed to control wirelessly the benchmark system, even though the incoming outputs from the slave board are lost. However, a buffered structure is implemented on the slave board to compensate the input losses of the arrived control sequences. The performance of the proposed wireless-networked predictive control compensation strategy for packet loss and perturbation handling in the wireless-networked control system in this work is verified through different experimentation conditions. Also, a comparative study with a wireless-networked proportional integral controller is performed to demonstrate the effectiveness of wireless-networked predictive control strategy for practical Internet of Things applications.

This article investigates the problem of data-driven cooperative tracking for a class of multi-agent linear systems under imperfect wireless communication. An autonomous Internet of Things predictive control application is designed to drive a robot with one wheel. The proposed methodology has been developed using Revolutionary Internet of Things Operating System running on STM32 and radio frequency communication shields over the User Datagram Protocol. To evaluate the performance of the predictive control algorithm, the User Datagram Protocol has been used due to the high number of packet losses in the communication channel. A robust analysis of Internet of Things technology among agents, combined with a network predictive control strategy against packet loss, limited bandwidth and attack links is carried out. The main feature of this methodology is that it is possible to achieve consensus monitoring and stability of closed-loop control systems. The efficiency of the proposed design approach is demonstrated by several experimental scenarios.

This paper uses the linear matrix inequality dilation approach to deal with robust stability and H∞ dynamic output feedback controller synthesis for linear parameter varying delayed systems with variable delay. This approach can express the original non-convex problem in terms of convex linear matrix inequalities and consequently reduces the conservatism of linear matrix inequality synthesis without dilation. Both delay-dependent stability and H∞ performance are studied in a quadratic context. Furthermore, a Lyapunov–Krasovskii functional is used to derive a delay-dependent criterion formulated in terms of a linear matrix inequality that will be used to search for an H∞ linear parameter varying delayed dynamic output feedback controller. To achieve this aim we use an integral inequality which plays a key role in the derivation of this criterion and enables the reduction of the H∞ cost in comparison to other results.

Advances in microelectronics and software allow the implementation of some control laws in embedded systems. In this work, the STM32 microcontroller is used to implement a robust fixed low-order controller for linear systems. The parametric uncertainty model is employed to describe the system behavior and the controller objective is to ensure, in the presence of model uncertainties and disturbances at the plant output, some specified time response performances. The controller design is expressed as a min–max non-convex optimization problem while taking into consideration uncertainties and the desired closed-loop specifications. Consequently, the use of a local optimization method to resolve such kind of problem may lead to a local solution and then, the obtained control law is not optimal. Therefore, with the purpose of obtaining an optimal solution which will be able to satisfy the desired specifications, the application of a global optimization method is recommended. The optimization method exploited in...

In this paper, we propose a feedforward neural networks-based robust predictive controller for a class of multi-input–multi-output non-linear systems. Using the structured uncertainties of the output layer’s weights of the neural networks model, the non-linear model of the real system is determined at each operating point. The control law is formulated as a minimax problem, which is solved online. The non-convex optimization is developed by minimizing the worst case of the objective cost function, taking into account the uncertainties of the non-linear model and the input control signal constraints. The efficiency of the proposed neural predictive controller is illustrated, in simulation, with a multivariable system example.