Real-time Control Experiments Research Articles

Deep Learning based Model Predictive Controller on a Magnetic Levitation Ball System

The magnetic levitation (maglev) ball system is a prototypical Single-Input-Single-Output (SISO) system, characterized by its pronounced nonlinearity, rapid response, and open-loop instability. It serves as the basis for many industrial devices. For describing the dynamics of the maglev ball system precisely in the pseudo linear model, the long short-term memory (LSTM) based auto-regressive model with exogenous input variables (LSTM-ARX) is proposed. Firstly, the LSTM network is modified by incorporating the auto-regressive structure with respect to sequence input, allowing it to deduce a locally linearized model without the need for Taylor expansion. Then, the LSTM-ARX model is transformed into a linear parameter varying (LPV) state space model, and upon this foundation, a model predictive controller (MPC) is proposed. Specifically, when deducing the MPC, the deep learning-based model is linearized by fixing its state input at the current state, so that the nonlinear, non-convex optimization problem can be converted to a finite-horizon quadratic programming problem, thereby deriving the explicit form of MPC. To further enhance the efficiency of the controller in real-time control tasks, a predictive functional controller (PFC) is proposed. It employs multiple nonlinear functions to fit the control sequence, thereby reducing the number of decision variables of the on-line optimization problem in MPC. The proposed controller was successfully applied to the real-time control of the maglev ball system. Simulation and real-time control experiments have validated the improvement in transient performance and efficiency of the LSTM-ARX model-based PFC (LSTM-ARX-PFC).

Real-time grid detection in sheet metal fiber laser cutting through coaxial monitoring

Real-time defects’ estimation and control of the cut quality through the coaxial camera monitoring of the kerf are amongst the most promising developments for laser cutting. In industrial systems, sheets are positioned on a metallic grid creating discontinuities in the cutting process due to unpredictable thickness, blown and resolidified material and the time varying position of the grid. The interaction between laser radiation escaping from the kerf and the grid, along with the difficult outflow of molten material, causes changes in the process emission images. These changes influence the defect estimation approach, and, consequently, the control action. In this work, a real-time grid identification and classification-based machine learning algorithm was developed and tested during the fusion cutting of 6 mm thick Al5754, exploiting a NIR coaxial monitoring system. Real-time control experiments with dross estimates were performed, demonstrating a correct identification of the grid and highlighting a feasible industrial application.

Comparative Study of Furuta Pendulum Based on LQR and PID Control

Furuta pendulum is an underdriven and unstable controlled object with one input and two outputs, and has an open-loop zero point at the origin of the complex plane. In this paper, a nonlinear dynamic model of the Furuta pendulum is established by the Eulerian-Lagrangian method, and a series-connected PID control strategy is designed by using the root locus method and the frequency domain analysis method based on the local linearization model at the steady-state operating point. Besides, an LQR control strategy is designed based on the discrete state-space model. In addition, a comparative analysis of the PID control strategy and the LQR control strategy is conducted, and it is found that the LQR control strategy can achieve stable control of the outer loop in the case of unstable inner loop. Finally, the effectiveness of the two control schemes is verified by real-time control experiments.

Compact high repetition rate Thomson parabola ion spectrometer.

We present the development of a compact Thomson parabola ion spectrometer capable of characterizing the energy spectra of various ion species of multi-MeV ion beams from >1020W/cm2 laser produced plasmas at rates commensurate with the highest available from any of the current and near-future PW-class laser facilities. This diagnostic makes use of a polyvinyl toluene based fast plastic scintillator (EJ-260), and the emitted light is collected using an optical imaging system coupled to a thermoelectrically cooled scientific complementary metal-oxide-semiconductor camera. This offers a robust solution for data acquisition at a high repetition rate, while avoiding the added complications and nonlinearities of micro-channel plate based systems. Different ion energy ranges can be probed using a modular magnet setup, a variable electric field, and a varying drift-distance. We have demonstrated operation and data collection with this system at up to 0.2Hz from plasmas created by irradiating a solid target, limited only by the targeting system. With the appropriate software, on-the-fly ion spectral analysis will be possible, enabling real-time experimental control at multi-Hz repetition rates.

Data-Driven Adaptive Optimal Control Under Model Uncertainty: An Application to Cold Atmospheric Plasmas

Cold atmospheric plasmas (CAPs) are increasingly used for applications requiring the processing of heat- and pressure-sensitive (bio)materials. A key challenge in model-based control of CAPs arises from the high-computational requirements of theoretical plasma models as well as lack of mechanistic understanding of plasma-surface interactions. Thus, control strategies that rely on simple, physics-based models that can be adapted to mitigate plant-model mismatch will be particularly advantageous for CAP applications. This article presents an optimal control approach for controlling the nonlinear and cumulative effects of CAPs delivered to a target surface using a simple system model. Through parsimonious input parameterization, the solution to the optimal control problem (OCP) is given by an arc sequence that does not include any singular arcs. A data-driven adaptive algorithm based on modifier adaptation is proposed to deal with the structural plant-model mismatch by estimating the mismatch in the cost and constraints of the OCP. The adaptive approach is shown to converge to a Karush–Kuhn–Tucker (KKT) point of the OCP for the true system. Moreover, a control strategy based on feedback linearization and derivative estimation is proposed for online tracking of path constraints in the presence of disturbances and model uncertainty. The proposed approach is demonstrated by simulations and real-time control experiments on a kilohertz-excited atmospheric pressure plasma jet in helium, in which the plasma treatment time is minimized while delivering a desired amount of nonlinear thermal effects to the target surface.

Enhancing Engine Order Sound Using Additive Feedforward Control for a Secondary Path with Uncertainty at Higher Frequencies

An active sound profiling (ASP) control algorithm based on the command FxLMS with an additive feedforward enhancement is suggested to overcome the limited performance to track a pre-defined target sound in a car cabin when an electronic sound generator (ESG) attached to a cowl panel of a car was adopted as a secondary actuator. As the uncertainty of the secondary path including the ESG-cowl pair was increased especially at higher frequencies, the tracking performance of a pure ASP algorithm was limited. The feedforward enhancement was added to the pure ASP algorithm to allow a robust tracking performance against a pre-defined target sound to enrich the insufficient engine sound at higher engine orders. After implementing this additive ASP approach in a test car, a real-time control experiment was carried out to demonstrate its tracking performance. A target sound was defined to cancel three engine order noises and to enhance the other three order sounds in the experiment. The experiment results showed that the proposed approach was able to provide an improved robust tracking performance at the engine orders for enhancing sound at higher frequencies compared to the pure ASP algorithm.

A Feasibility Study of an ESG to Suppress Road Noise of a Car

This study considered implementing an active road noise control (ARNC) system using an electronic sound generator (ESG) as a secondary actuator to suppress road noise in a car cabin. The ESG was installed to the cowl panel of a test car to generate structure-borne anti-noise by vibrating the panel. A robust multiple-reference single-input single-output (MR-SISO) ARNC algorithm based on the FxLMS was designed. Four 3-axis accelerometers and a microphone were adopted to acquire the reference signals and the error signal for the control algorithm. The radiated sound pressure from the ESG–cowl pair was high enough to suppress the road noise at a car speed of 60 kph. The optimized least number of reference signals and their locations were determined after computer simulation from the measured primary path data. Real-time control experiments showed an A-weighted sound pressure level reduction of 6.0 dB in the average of three dominant road booming noises in 100–250 Hz with the four optimized reference signals at 60 kph. More reference signals gave a further reduction such as 8.3 dB with 12 reference signals. Thus, this study suggests that the ESG coupled with the cowl panel can be an affordable alternative as a secondary actuator in an ARNC system to suppress road noise in a car.

A reinforcement learning-based near-optimal hierarchical approach for motion control: Design and experiment

As a data-driven design method, model-free optimal control based on reinforcement learning provides an effective way to find optimal control strategies. The design of model-free optimal control is sensitive to system data because it relies on data rather than detailed dynamic models. A prerequisite for generating applicable data is that the system must be open-loop stable (with a stable equilibrium point), which restricts the data-based control design methods in actual control problems and leads to rare experimental studies or verification in the literature. To improve this situation and enrich its applications, we propose a pre-stabilized mechanism and apply it to the motion control of a mechanical system together with a reinforcement learning-based model-free optimal control method, which constitutes a so-called hierarchical control structure. We design two real-time control experiments on an underactuated system to verify its effectiveness. The control results show that the proposed hierarchical control is quite promising in controlling this mechanical system, even though it is open-loop unstable with unknown dynamics.

Real-time analysis of adaptive fuzzy predictive controller for chaotification under varying payload and noise conditions

In this paper, an adaptive fuzzy generalized predictive controller is proposed for the anti-chaos control where the system dynamics are assumed to be unknown or uncertain in large. In the proposed mechanism, the control signal is generated by an input constraint generalized predictive controller to track the Duffing oscillator which is used as a chaotic reference system. The adaptive fuzzy system, as a real plant model, is employed to obtain prediction through the receding horizon. In order to show the capability of the proposed mechanism, three different payloads and noise cases are studied in the real-time control experiments of flexible-joint manipulator and tracking results are compared with three more non-model-based controllers which are conventional adaptive fuzzy control, indirect adaptive fuzzy sliding mode controller and proportional–integrative–derivative control. Results show that the proposed mechanism has better control performance than the others.

Analysis and variable step control of a bidirectional complementary-type inchworm actuator

To meet the demands of high speed and large loading capacity in various nanopositioning applications, a voltage sequence based step analysis and the corresponding variable step control strategy are explored for a bidirectional complementary-type inchworm actuator. Specifically, the bidirectional variable step motion can be achieved without significant drop of the loading capacity by controlling the two-channel input voltage sequences, where the variable speed motions can help to achieve more rapid nano-positioning tasks. In particular, we introduce the voltage sequences in the modelling of the step length and the total clamping force for the purpose of achieving controllable bidirectional motion with large loading capacity. Real time control experiments are comprehensively conducted, which demonstrates a bidirectional variable step motion range of ±22 µm is achieved by controlling two-channel input signals. In addition, the actuator with the proposed method has an adjustable total clamping force capable of improving the motion stability and loading capacity in variable step motion, which are favourable to various engineering applications in high speed nanopositioning areas.

Data-driven LPV model predictive control of a cold atmospheric plasma jet for biomaterials processing

Cold atmospheric plasmas (CAPs) are increasingly used for treatment of complex surfaces in biomedical and biomaterials processing applications. However, the multivariable, distributed-parameter, and nonlinear nature of CAP dynamics and plasma–surface interactions, coupled with the sensitivity of plasmas to exogenous disturbances, make their safe, reproducible and effective operation challenging. This paper adopts a data-driven linear parameter-varying (LPV) modeling framework to learn a control-oriented model for predictive control of a kHz-excited atmospheric pressure plasma jet in Helium. A hierarchical model-based control strategy is proposed based on a supervisory LPV-based model predictive controller (LPV–MPC) to regulate the nonlinear thermal effects of plasma on a surface. Real-time control experiments demonstrate the effectiveness of the proposed LPV–MPC strategy for the multivariable control of surface temperature and plasma optical intensity, as well as for controlling the spatial delivery of the cumulative thermal effects of the plasma jet on a surface.

Robust predictive control of coupled water tank plant

Liquid level control of the liquid tank is a basic control problem in the process industry. The liquid level control system usually has the characteristics of strong coupling, nonlinearity and large lag. Up to now, PID control is the main method of liquid level control in industry, but it is difficult to obtain high precision control performance. The difficulty of model-based liquid level control such as Robust Model Predictive Control (RMPC) is hard to obtain an accurate model of the plant or need accurate steady-state information which is hard to be obtained in practice. This paper provides a systematic modeling and RMPC method to implement a Coupled Water Tank (CWT) plant output-tracking control in the case without whole steady-state information of the system. First, a data-driven modeling technique based on the Radial Basis Function (RBF) net-type coefficients Multi-Input/Multi-Output (MIMO) AutoRegressive model with eXogenous variable (ARX) is applied to build the RBF-ARX model of the CWT plant. Using the model, a locally linearized model and a polytopic uncertain linear parameter varying (LPV) model are constructed to represent the current and future behavior of the nonlinear CWT system. Based on the two models, an RMPC method is designed to implement the system’s output tracking control under the condition without whole steady-state knowledge of the CWT plant. The stability of the RMPC strategy is guaranteed by using the time-varying parameter-dependent Lyapunov function and feasibility of the Linear Matrix Inequalities (LMIs). Real-time control experiments and comparison with other control methods are carried out on the CWT system, and the results show that the presented method is superior.

Optimal controller applied to robotic systems using covariant control equations

ABSTRACT We elaborate an algorithm to optimally control the actuators of robot manipulators. By regarding the robot kinetic energy as a metric, the dynamic model is formulated using tensors in a Riemannian geometry manifold. The control criterion is selected to be invariant, leading to covariant control equations that take advantage of the system dynamics. Motion equations are associated with conjugate control equations resulting in a system of second order ordinary differential equations. The suggested control method consists in solving this system of equations to obtain optimal joint forces off-line and use them as inputs in a control scheme to track optimal trajectories. Position-velocity feedback ensures proper behaviour. Motion control simulations show that the optimal trajectory can be tracked with good accuracy, even with imposed perturbations. A stability analysis for switched systems precedes real-time motion control experiments conducted on a serial robot manipulator with revolute joints to validate our method.

Adaptive Algorithm of a Tap-Changer Controller of the Power Transformer Supplying the Radial Network Reducing the Risk of Voltage Collapse

The development of renewable energy, including wind farms, photovoltaic farms as well as prosumer installations, and the development of electromobility pose new challenges for network operators. The results of these changes are, among others, the change of network load profiles and load flows determining greater volatility of voltages. Most of the proposed solutions do not assume a change of the transformer regulator algorithm. The possibilities of improving the quality of regulation, which can be found in the literature, most often include various methods of coordination of the operation of the transformer regulator with various devices operating in the Medium-Voltage (MV) network. This coordination can be decentralized or centralized. Unfortunately, the proposed solutions often require costly technical resources and/or large amounts of real-time data monitoring. The goal of the authors was to create an algorithm that extends the functionality of typical transformer control algorithms. The proposed solution allows for reducing the risk of voltage collapse. The performance of the proposed algorithm was validated using multivariate computer simulations and tests with the use of a physical model of the distribution network. The DIgSILENT PowerFactory environment was used to develop the simulation model of the proposed algorithm. Then, tests were conducted on real devices installed in the LINTEˆ2 Laboratory at the Gdańsk University of Technology, Poland. Selected test results are included in this paper. All results have shown that the proposed algorithm makes it possible to increase the reserve of the voltage stability of the node, in which it is applied, thus mitigating the risk of a voltage collapse occurring. The proposed algorithm does not require complex and costly technical solutions. Owing to its simplicity, it has a high potential for practical application, as confirmed by the real-time control experiment in the laboratory.

Myoelectric digit action decoding with multi-output, multi-class classification: an offline analysis

The ultimate goal of machine learning-based myoelectric control is simultaneous and independent control of multiple degrees of freedom (DOFs), including wrist and digit artificial joints. For prosthetic finger control, regression-based methods are typically used to reconstruct position/velocity trajectories from surface electromyogram (EMG) signals. Unfortunately, such methods have thus far met with limited success. In this work, we propose action decoding, a paradigm-shifting approach for independent, multi-digit movement intent prediction based on multi-output, multi-class classification. At each moment in time, our algorithm decodes movement intent for each available DOF into one of three classes: open, close, or stall (i.e., no movement). Despite using a classifier as the decoder, arbitrary hand postures are possible with our approach. We analyse a public dataset previously recorded and published by us, comprising measurements from 10 able-bodied and two transradial amputee participants. We demonstrate the feasibility of using our proposed action decoding paradigm to predict movement action for all five digits as well as rotation of the thumb. We perform a systematic offline analysis by investigating the effect of various algorithmic parameters on decoding performance, such as feature selection and choice of classification algorithm and multi-output strategy. The outcomes of the offline analysis presented in this study will be used to inform the real-time implementation of our algorithm. In the future, we will further evaluate its efficacy with real-time control experiments involving upper-limb amputees.

An intention-based online bilateral training system for upper limb motor rehabilitation

Bilateral rehabilitation training robotic systems have potential to promote the upper limb motor recovery of post-stroke hemiparesis patients through providing the synchronization motion between the impaired limb and contralateral limb. The active rehabilitation training based on patients’ intention can also promote the recovery effect by stimulating the activity of the ipsilateral hemisphere and contralateral hemisphere. In this paper, a novel intention-based bilateral training system using biomedical signals which represents the muscle activity information and active motion intention was proposed to promote the rehabilitation training effect. The proposed system can provide the synchronization motion to the impaired limb by the exoskeleton device according to the sEMG signals from the contralateral intact limb. A BPNN model using a novel multi-features input vector was employed for establishing the relationship between the sEMG signals and the motion intention. To verify the intention prediction performance, the comparison experiments involving both the offline phase and online phase were carried out using three different kinds of feature input vectors of sEMG. Furthermore, the real-time bilateral control experiments were conducted to verify the feasibility and effectiveness of the proposed bilateral rehabilitation system, in terms of motion synchronization tracking and the real-time characteristics.

A Real-Time Image Encryption Method of Networked Inverted Pendulum Visual Servo Control System

In the real-time control of networked inverted pendulum visual servo control system (NIPVSCS), how to achieve safe and efficient image transmission is an important and challenging problem. Generally, the inverted pendulum images are directly transmitted via the network, while these images are not secure because they may be attacked by the hacker. To solve the problem, a fast image encryption method by combining image scaling and improved self-diffusion image encryption (ISDIE) algorithm is firstly proposed, which is employed to encrypt the captured images after they are scaled. The security performance of the proposed algorithm, the characteristics of different delays, and the control performance of the NIPVSCS are then analyzed. Finally, simulation and real-time control experiments confirm the feasibility and effectiveness of the proposed method.

Research on Visual Control System of Inverted Pendulum Based on Pixel Displacement

In inverted pendulum visual servo control system (IPVSCSs), how to achieve the fast and effective measurement of the state information of the inverted pendulum is an important task. In this paper, the characteristics of the inverted pendulum are considered and a series of image processing and positioning algorithm of the pendulum are proposed. Furthermore, the inverted pendulum control system model based on the pixel displacement is established and the LQR controller is designed. Finally, the feasibility and superiority of the proposed method are verified by real-time control experiments.

Multiple model hybrid adaptive vibration control for flexible cantilever beam with varying load

To solve the vibration suppression problem for a piezoelectric flexible cantilever beam with varying load, this study proposes a new type of multiple model hybrid adaptive vibration control method with a new multiple model switching cost index function. It employs a pre-filter consisting of secondary path and positive feedback path model to eliminate the positive feedback between the secondary actuator and reference sensor. By constructing a piezoelectric flexible cantilever beam as the controlled plant, we constructed a real-time control verification experiment platform based on a real-time MATLAB xPC target. We did a real-time control experiment comparison for vibration suppression of single-frequency, narrowband, and broadband disturbances. Experimental results show that the proposed multiple model hybrid adaptive vibration control algorithm is feasible and has good vibration suppression performance with rapid convergence speed.

A GPC-based Multi-variable PID Control Algorithm and Its Application in Anti-swing Control and Accurate Positioning Control for Bridge Cranes

It is one of the key tasks for the bridge crane to achieve anti-swing control of the hook and the accurate positioning of the body to work efficiently, safely and automatically. Based on Lagrange equation, this paper is to propose a dynamic model of the crane motion system for designing controller. In the controller design, Proportional-Integral-Derivative (PID), the most widely used controller in engineering, is adopted and a new parameter tuning algorithm for a multi-variable PID controller based on generalized predictive control (GPC) is given. It is found that the multi-variable PID controller shares the same structural mathematical expressions with the GPC law, which makes for the transfer and calculation of the three parameters P, I and D, and that the new algorithm enables the traditional PID controller to perform as brilliantly as the GPC. The results of both the simulation and real-time control experiments show that the newly-proposed PID controller can effectively eliminate the swing of the hook and control the bridge cranes moving position accurately.