Data-driven Control Research Articles

Data-driven and autonomous manufacturing control in cyber-physical production systems

Modern manufacturing networks consist of cyber-physical systems (CPS) which offer an array of interesting capabilities, ranging from local computation over data generation to communication capabilities. As traditional control approaches fail to fully leverage these capabilities, the last decade has seen a renewed interest in distributed control approaches based on autonomous entities. In this article, we study the synergistic potentials of autonomous control and machine learning in a job-shop setting, addressing challenges of modern manufacturing such as market fluctuation and process time variance, thus leveraging the potentials of CPS in order to flexibly configure manufacturing networks and achieve cost-minimal production. We utilize a multi-agent based discrete-event simulation to compare this novel approach to a traditional heuristic, underlining the potentials and advantages of data-driven control approaches.

Learning of model-plant mismatch map via neural network modeling and its application to offset-free model predictive control

We propose an improved offset-free model predictive control (MPC) framework, which learns and utilizes the intrinsic model-plant mismatch map, to effectively exploit the advantages of model-based and data-driven control strategies and overcome the limitation of each approach. In this study, the model-plant mismatch map on steady-state manifold is approximated via artificial neural network (ANN) modeling based on steady-state data from the process. Though the learned model-plant mismatch map can provide the information at the equilibrium point (i.e., setpoint), it cannot provide model-plant mismatch information during transient state. To handle this, we additionally apply a supplementary disturbance variable which is updated from a revised disturbance estimator considering the disturbance value obtained from the learned model-plant mismatch map. Then, the learned and supplementary disturbance variables are applied to the target problem and finite-horizon optimal control problem of the offset-free MPC framework. By this, the control system can utilize both the learned model-plant mismatch information and the stabilizing property of the nominal disturbance estimator. The closed-loop simulation results demonstrate that the proposed offset-free MPC scheme utilizing the model-plant mismatch map learned via ANN modeling efficiently improves the closed-loop reference tracking performance of the control system. Additionally, the zero-offset tracking condition of the developed framework is mathematically examined.

Stochastic Learning of Energy System for Data-Driven Control in Manufacturing Process

Stochastic Learning of Energy System for Data-Driven Control in Manufacturing Process

Gain-scheduled Control with Data-driven Control and Model-matching Method for Vehicle

Gain-scheduled Control with Data-driven Control and Model-matching Method for Vehicle

Data-Driven Fast Voltage Control in Non-DPMU Distribution Networks With Microgrids

Traditional voltage control methods for distribution networks assume perfect knowledge of the power system model. Nevertheless, the extensive scale of future distribution networks makes it unrealistic to acquire the overall operation state monitoring. Moreover, with the deregulation of distribution networks, partial controllable resources belong to independent systems, such as microgrids, causing distribution system operators unable to force them to provide voltage support directly. To cope with the previously mentioned problems, a data-driven fast voltage control method for distribution networks with MGs is proposed in this article. First, voltage sensitivity matrices are estimated indirectly by identifying line parameters in a regression approach, without using measurement data of distribution phasor measurement units (DPMUs) in distribution networks. Then, an incomplete information game model is proposed to motivate MGs to provide ancillary services of voltage control. To guarantee privacy, only a little key information is shared among MGs and distribution system operators. Moreover, MGs make voltage control strategies autonomously based on the data-driven deep reinforcement learning algorithms, while maximizing their own profits. Finally, we test the method on the modified IEEE 33-node networks and IEEE 123-node networks. The results demonstrate that the proposed method can provide an accurate voltage estimation in electricity markets with non-DPMU measurement data and increase energy and asset utilization.

Data-driven fault-tolerant control for unmanned aerial vehicles without using identification model

Unmanned aerial vehicle (UAV)’s fault-control problem was studied in this paper, and data-driven fault-tolerant control scheme was developed for acceleration tracking control of UAV in order to cope with the uncertainties induced by aerodynamic damage. A linear UAV dynamic model was given with reasonable assumptions, and the acceleration tracking control for UAV was converted to solving an infinite-horizon optimal control problem. The augmented algebraic Riccati equation (ARE) is derived, and its solution stability is proved based on Lyapunov theory. The data-driven control algorithm is further derived for online solving of the augmented ARE with only using flight data. The proposed algorithm is based on experience replay of flight data rather than model knowledge, so it greatly reduces the effect of uncertainties induced by aerodynamic damage on the flight control system for UAVs. Finally, the effectiveness of developed algorithm is verified through the numerical simulations under different uncertainties induced by aerodynamic damage.

Closed-Loop Dynamic Control of a Soft Manipulator Using Deep Reinforcement Learning

The focus of the research community in the soft robotic field has been on developing innovative materials, but the design of control strategies applicable to these robotic platforms is still an open challenge. This is due to their highly nonlinear dynamics which is difficult to model and the degree of stochasticity they often incorporate. Data-driven controllers based on neural networks have recently been explored as a viable solution to be employed for these manipulators. This letter presents a neural network-based closed-loop controller, trained by a deep reinforcement learning algorithm called Trust Region Policy Optimization (TRPO). The training takes place in simulation, using an approximation of the robot forward dynamic model obtained with a Long-short Term Memory (LSTM) network. The trained controller allows following different paths executed with different velocities in the workspace of the robot. The results demonstrate that the controller is effective in normal working conditions and with a payload attached to the end-effector of the manipulator.

Towards a generic physics-based machine learning model for geometry invariant thermal history prediction in additive manufacturing

• Alternative to the timely and costly finite element-based simulations of the thermal history in additive manufacturing. • Generic, physics-based features that performs well on various geometries, deposition patterns, meshes, and other simulation properties. • High precision machine learning predictions of thermal fields in additive manufacturing that achieved mean absolute percentage error below 5%. Additive manufacturing (AM) is an emerging manufacturing technology that constructs complex parts through layer-by-layer deposition. The prediction and control of thermal fields during production of AM parts are of crucial importance as the temperature distribution and gradient dictates the microstructures, properties, and performance. Finite element (FE) analyses are commonly conducted to simulate the thermal history of the AM process, but are known to be costly and time-consuming. This paper aims to address the challenge by presenting the essential components of a generic data-driven control framework. The proposed framework utilizes extremely randomized trees and is trained and tested on datasets generated through FE simulations. The datasets contain generic, engineered features constructed based on the physics of the underlying thermal process. The features are transferable between a wide range of cases and have achieved mean absolute percentage errors (MAPE) below 2.5% for predicting nodal temperature profiles. In addition, predictions of entire simulations with machine learning (ML) models trained on datasets from different cases have been conducted with MAPE below 5%. The results demonstrate the transferability of thermal histories between several geometries and significantly reduce the need for expensive FE simulations. We believe that these findings are an important step towards real-time optimization in AM.

Data-Driven Control for Radiative Collapse Avoidance in Large Helical Device

Data-Driven Control for Radiative Collapse Avoidance in Large Helical Device

Hybrid Decision Based on DNN and DTC for Model Predictive Torque Control of PMSM

To address the issue of poor real-time performance caused by the heavy computational burden of the finite control set model predictive torque control (MPTC) of a permanent magnet synchronous motor (PMSM), a data-driven control method using a deep neural network (DNN) is proposed in this paper. The DNN can learn the MPTC’s selective laws from its operation data by training offline and then substitute them for voltage vector selection online. Aiming to address the data-driven runaway problems caused by the asymmetry between the dynamic and static training data, a hybrid decision control strategy based on DNN and DTC (direct torque control) is further proposed, which can realize four-quadrant operation with a control effect basically equivalent to MPTC. The proposed strategy has great application potential for use in multi-level inverter and matrix converter driving with multiple candidate voltage vectors or multi-step prediction.

Data-Driven Cooperative Control Model of Shearer-Scraper Conveyor Based on Rough Set Theory

The cooperative control of shearer and scraper conveyors is the prerequisite for the realization of intelligent comprehensive mining equipment and unmanned comprehensive mining workings. However, because of the harsh working face environment, the complex process of comprehensive mining, and the many uncertainties, it is difficult to establish a mathematical model for the cooperative control of shearer and scraper conveyors precisely through the operating mechanism. In the era of big data, the data-driven model has become a popular trend. Therefore, according to the actual production process data, this article proposed a data-driven cooperative control model of shearer–scraper conveyor based on rough set theory. First, the selection method of process monitoring parameters based on rough set theory was proposed to remove redundant parameters and redundant parameter values. Moreover, the decision rule base of cooperative speed regulation of shearer and scraper conveyor was established. Then a collaborative speed regulation decision algorithm based on attribute importance was designed. The algorithm matches the decision rules according to the real-time observation data and then determines the running speed of the shearer. The simulation results show that the proposed data-driven collaborative control model of shearer–scraper conveyor based on rough set theory overcomes the limitations of the mathematical model. It can predict the running speed of shearer well and realize the collaborative speed regulation of shearer–scraper conveyor.

Data-Driven Method for Predicting Soil Pressure of Foot Blades within a Large Underwater Caisson

The soil pressure on the bottom surface of the foot blades is an important monitoring point during the sinking process of large underwater caissons. Complex soil-structure interactions occur during the sinking process, making it difficult to accurately predict the soil pressure of foot blades. Accurate construction processes often rely on data from the soil pressure of foot blades in the field. In this study, a data-driven approach is used to establish the relationship between the amount of sinking of the caisson and the soil pressure of foot blades. Furthermore, by improving the splitting method of the original Classification and Regression Tree (CART) algorithm, a single model’s numerical prediction of 80-foot blades soil pressures is realized. The improved CART model, multilayer perceptron (MLP), long short-term memory (LSTM), and a linear regression model are compared through a comprehensive multiparameter evaluation method. Finally, this article discusses the deployment scheme of the model by comparing and analyzing the data in the time period of 10 : 00 on July 29, 2020, and 23 : 00 on August 7, 2020. The experimental results can satisfy the engineering demands and provide a basis for further data-driven intelligent control of large caisson sinking.

Direct data-driven control approach of reference shaping for two degree of freedom control

Direct data-driven control approach of reference shaping for two degree of freedom control

Data-driven control of the coupled loads for floating wind turbine systems based on deformable trailing edge flaps

Load controller design for floating wind turbine systems (FWTS) poses unique challenges owing to the couplings of platform motions, blade loads and rotor power. Any controller based on a single objective is likely to worsen other performance metrics. This paper presented a data-driven control system to comprehensively handle that muti-objective coupling problem. The control system was realized by introducing a model free adaptive control (MFAC) based on deformable trailing edge flap (DTEF) and a feedback compensation in the blade pitch control loop. The identification error and actuator saturation were considered while designing the control law. The proposed control system was numerically tested in the modified FAST code equipped with DTEF interfaces. Numerical results showed considerable load reductions on blades and floating platforms at the cost of the added DTEF actuation and a minor increase in blade pitch rate. Moreover, the generator power fluctuations and the fatigue loading on drive train and tower were also alleviated. In addition, the synchronous aero-hydro-elastic coupling relationships in the wind-blade-wave-platform system were impaired by introducing the servo-dynamics of DTEF and blade pitch feedback compensation. Hence the total energy flowing into FWTS was dissipated.

Pseudo-spectral optimization and data-driven control of vehicle start process with dual-clutch transmission

In this study, a dual-clutch transmission start control strategy based on pseudo-spectral optimization and data-driven control is proposed to respond to the time-varying start intentions, to reduce friction and jerk, and to improve the start quality. First, taking the jerk, start time, and friction work as the optimization indexes, the adaptive pseudo-spectral method is used to accurately determine the optimal clutch engagement target trajectory. Then, a real-time planning method for the target clutch engagement trajectory is proposed based on a dual deep gated recurrent unit network. Second, through partial least squares identification, a data-driven predictive control (DDPC) method based on the autoregressive with exogenous model is proposed. The model parameters are updated online by a multi-model filtering algorithm based on input and output data, allowing real-time adaptation to the system uncertainty and nonlinear characteristics. Simulations and experiments show that the proposed target clutch trajectory planning method can effectively respond to time-varying start intentions in real-time. In addition, the DDPC controller can adapt to the uncertainty and nonlinear characteristics of the system and, hence, improve the control accuracy.

Input torque measurements for wind turbine gearboxes using fiber-optic strain sensors

Abstract. Accurate knowledge of the input torque in wind turbine gearboxes is key to improving their reliability. Traditionally, rotor torque is measured using strain gauges bonded to the shaft. Transferring the resulting signal from the rotating shaft to a stationary data acquisition system while powering the sensing devices is complex and costly. The magnitude of the torques involved in wind turbine gearboxes and the high stiffness of the input shaft pose additional difficulties. This paper presents a new alternative method to measure the input torque in wind turbine gearboxes based on deformation measurements of the static first-stage ring gear. We have measured deformation using fiber-optic strain sensors based on fiber Bragg gratings because of their advantages compared to conventional electrical strain gauges. The present study was conducted on a Siemens Gamesa Renewable Energy gearbox with a rated power of 6 MW, in which a total of 54 fiber-optic strain sensors were installed on the outer surface of the first-stage ring gear. The gear mesh forces between the planets and the ring gear cause measurable deformations on the outer surface of the stationary ring gear. The measured strains exhibit a dynamic behavior. The strain values change depending on the relative position of the strain sensors to the planet gears, the instantaneous variations of the input torque, and the way load is shared between planets. A satisfactory correlation has been found between the strain signals measured on the static ring gear and torque. Two signal processing strategies are presented in this paper. The first procedure is based on the peak-to-peak strain values computed for the gear mesh events, and therefore torque can only be estimated when a gear mesh event is detected. The second signal processing procedure combines the strain signals from different sensors using a Coleman coordinate transformation and tracks the magnitude of the fifth harmonic component. With this second procedure, it is possible to estimate torque whenever strain data of all sensors are available, leading to an improved frequency resolution up to the sampling frequency used to acquire strain data. The method presented in this paper could make measuring gearbox torque more cost-effective, which would facilitate its adoption in serial wind turbines and enable novel data-driven control strategies, as well as a more accurate assessment of the consumed fatigue life of the gearboxes throughout their operation.

Remarks on input to state stability of perturbed gradient flows, motivated by model-free feedback control learning

Recent work on data-driven control and reinforcement learning has renewed interest in a relative old field in control theory: model-free optimal control approaches which work directly with a cost function and do not rely upon perfect knowledge of a system model. Instead, an “oracle” returns an estimate of the cost associated to, for example, a proposed linear feedback law to solve a linear–quadratic regulator problem. This estimate, and an estimate of the gradient of the cost, might be obtained by performing experiments on the physical system being controlled. This motivates in turn the analysis of steepest descent algorithms and their associated gradient differential equations. This note studies the effect of errors in the estimation of the gradient, framed in the language of input to state stability, where the input represents a perturbation from the true gradient. Since one needs to study systems evolving on proper open subsets of Euclidean space, a self-contained review of input to state stability definitions and theorems for systems that evolve on such sets is included. The results are then applied to the study of noisy gradient systems, as well as the associated steepest descent algorithms.

Human-Tailored Data-Driven Control System of Autonomous Vehicles

Path Tracking Controller (PTC) plays a key role in achieving improved dynamic behaviour of Autonomous Vehicle (AV) while it is mainly responsible for implementing the planned paths. In this work, a learning-based Model Predictive Controller (MPC) is proposed to integrate a data-driven approach to the path tracking task using human driving demonstrations. The objective is to reproduce customised vehicle motion profile that is statistically comparable to a human driver with a proper trade-off between vehicle speed-acceleration profile and tracking accuracy. A combined longitudinal and lateral MPC controller with a novel feature-based parametric cost function is designed to perform the path tracking task under different road conditions. A high precision data logging system is developed to collect human driving demonstration data. From the analysis of the collected data, relevant features are chosen based on their applicability to produce a customised motion profile while performing AV’s path tracking task. Next, a bi-level optimisation-based Inverse Optimal Controller (IOC) is designed to learn the cost function’s parameters using the human demonstration data. The outcomes of the study show that the proposed controller can effectively regenerate some characteristics of human driving, including speed, longitudinal and lateral accelerations, and yaw rate.

A convex optimization approach to synthesizing state feedback data-driven controllers for switched linear systems

This paper seeks to develop a computationally tractable framework for data-driven control of switched linear systems. Specifically, given a model structure and experimental data collected at different operating points, we seek to directly design a state-feedback controller that stabilizes a system that arbitrarily switches amongst all sub-systems that could have generated the observed data, without an explicit plant identification step. The main result of the paper shows that this robust optimization problem can be recast, through the use of duality, into a polynomial optimization form and efficiently solved, leading to a robust controller with guaranteed ℓ∞ worst-case performance. The effectiveness of the proposed technique is illustrated with several examples, including control of the horizontal motion of a quadcopter

Model-free repetitive control design and implementation for dynamical galvanometer-based raster scanning

Data-driven repetitive control (RC) is proposed in this work to track online, dynamical raster trajectories in galvanometer-based scanning. To remove the requirement of a plant model in conventional model-based RC, we use model-free iterative learning control (ILC) to synthesize the data-driven repetitive controllers. Specifically, the frequency-domain plant-inversion and loop-shaping methods are both converted into time-domain trajectory tracking problems. The ILC is then applied to solve the trajectory tracking problems and subsequently derive the repetitive controllers from data. The stability conditions of both methods are analyzed and used to guide the data-driven control design. Experimental results on a commercially available galvanometer scanner demonstrate that the proposed methods improve the tracking error of a predefined raster scan by more than 30 times, as the conventional ILC does. Moreover, after applying data-driven RC, users can online assign various center positions and magnitudes of the raster trajectory. Once assigning a new reference in this continuous mode, the tracking error rapidly converges to the steady-state within ten periods.