Direct Learning Control Research Articles

How could data integrate with control? A review on data-based control strategy

In recent years, data-driven research has attracted significant attention from both industry and academia due to its success in medical, transportation, finance and other fields. How could data tightly integrate with control is a hot issue being explored. This paper provides a brief survey on the relationship of data and control. Four types of data-based control studies are considered, namely post-identification control, direct data control, learning control, and observer-based control. The observer based control can provide real-time observation and estimation of the unknown dynamic systems. It has a great potential in solving complex online control problems. This paper classifies and analyzes the observer integrated control, covering the latest research progress and forecasting the development trend in this direction. This paper concludes with a summary advantages and disadvantages of data-based controls.

The Influence of Scientific Independence towards Students’ Content Analysis and Science Process Skills on Cell Metabolism Topic

Students commonly judge that cell metabolism topic is arduous since they only listen to their teacher and memorizing concept without a deep comprehension. Other than that, a direct explanation could not enhance the students’ science process skills. Thus, this research intended to analyze the differences of the content-analysis and science process skills between the students experiencing a learning that oriented to scientific independence (experimental class) and those joining a direct learning (control class). The materials taught in both classes were cell metabolism. This study employed the quantitative of the quasi-experimental method. The research object was the students learning Biology. The students joining a scientific independence-based class were able to answer questions requiring analytical thinking. The relationship of the students’ scientific independence towards the content analysis and science process skills was seen on the test. Based upon the statistical data analysis, the correlation significance was 0,038<0,05. Therefore, it concluded that there was a very significant correlation between the students’ scientific independence towards the content analysis and science process skills.

A boundedness result for the direct heuristic dynamic programming

Approximate/adaptive dynamic programming (ADP) has been studied extensively in recent years for its potential scalability to solve large state and control space problems, including those involving continuous states and continuous controls. The applicability of ADP algorithms, especially the adaptive critic designs has been demonstrated in several case studies. Direct heuristic dynamic programming (direct HDP) is one of the ADP algorithms inspired by the adaptive critic designs. It has been shown applicable to industrial scale, realistic and complex control problems. In this paper, we provide a uniformly ultimately boundedness (UUB) result for the direct HDP learning controller under mild and intuitive conditions. By using a Lyapunov approach we show that the estimation errors of the learning parameters or the weights in the action and critic networks remain UUB. This result provides a useful controller convergence guarantee for the first time for the direct HDP design.

Implementation of an Intelligent Learning Controller for Gait Control of Biped Walking Robot

This paper presents an intelligent learning controller for repetitive walking motion of biped walking robot. The proposed learning controller consists of an iterative learning controller and a direct learning controller. In the iterative learning controller, the PID feedback controller takes part in stabilizing the learning control system while the feedforward learning controller plays a role in compensating for the nonlinearity of uncertain biped walking robot. In the direct learning controller, the desired learning input for new joint trajectories with different time scales from the learned ones is generated directly based on the previous learned input profiles obtained from the iterative learning process. The effectiveness and tracking performance of the proposed learning controller to biped robotic motion is shown by mathematical analysis and computer simulation with 12 DOF biped walking robot.

A Fuzzy Neural Network Direct Adaptive Iterative Learning Controller for Robotic Systems

Abstract This paper studies the iterative learning control of robotic systems with repetitive tasks. A fuzzy neural network is applied to design a direct adaptive iterative learning controller. The fuzzy neural network is introduced for compensation of the unknown certainty equivalent controller. A new adaptive law using mixed time-domain and iteration-domain adaptation is developed. It is shown that the finiteness of control parameters and control input can be guaranteed for all the time interval during each iteration without using parameter projection.

Supplementary Damping Controller Design using Direct Heuristic Dynamic Programming in Complex Power Systems

In modern, large scale interconnected power grids, low-frequency oscillation is a key roadblock to improved power transmission capacity. Supplementary generator control, flexible AC transmission system (FACTS), and high voltage direct currents (HVDC) are engineered devices designed to damp such low frequency swings. In this paper a neural network-based approximate dynamic programming method, namely direct heuristic dynamic programming (direct HDP), is applied to power system stability enhancement. Direct HDP is a learning and approximation based approach to addressing nonlinear system control problems under uncertainty, and it is also a model-free design strategy. The action and critic networks of the direct HDP are implemented using multi-layer perceptrons; learning is carried out based on the interactions between the controller and the power system. For this design approach, real time system responses are provided through wide-area measurement system (WAMS). The controller learning objective is formulated as a reward function that reflects global characteristics of the power system under low frequency oscillation, as well as tight coupling effects among system components. Direct HDP control design is illustrated by case studies, which are also used to demonstrate the learning control performance. The proposed direct HDP learning control is also developed as a new solution to a large scale system coordination problem by using the China Southern Power Grid as a major test bed.

Analysis and Design of Iterative Learning Control Strategies for UPS Inverters

In this paper, two iterative-learning-based control methods for uninterruptible power system (UPS) inverters, viz., direct iterative learning controller (direct ILC) and hybrid iterative learning controller (hybrid ILC), are proposed. In both methods, a zero-phase filter that is designed in the frequency domain is applied in order to provide compensation for the resonant peak in the system model to ensure error convergence. Furthermore, a "forgetting factor" is introduced in both control algorithms to increase the robustness of the scheme against measurement noise, initialization error, and/or variation of system dynamics due to any parameter drift. In the direct ILC method, ILC is combined with the feedforward of the reference to improve dynamic performance. This method is shown to be capable of achieving very high steady-state performance although its dynamic response was not very good. In the hybrid ILC method, ILC is combined with a proportional-derivative controller aside from the reference feedforward to improve the dynamic response further. Experimental results show that both proposed controllers can achieve very low total harmonic distortion and fast error convergence under different loads while using only one sensor. The proposed direct ILC is an effective solution for UPS products where high-quality steady-state output voltage is more important than fast dynamic response, while hybrid ILC can result in excellent steady-state performance with a more improved dynamic response than direct ILC

Direct learning control design for a class of linear time-varying switched systems

In this brief, a direct learning control method for a class of switched systems is proposed. The objective of direct learning is to generate the desired control profile for a newly switched system without any feedback, even if the system may have uncertainties. This is achieved by exploring the inherent relationship between any two systems before and after a switch. The new method is applicable to a class of linear time varying, uncertain, and switched systems, when the trajectory tracking control problem is concerned. A numerical simulation demonstrates the effectiveness of the proposed method.

Recursive Direct Learning of Control Efforts for Trajectories with Different Magnitude Scales

ABSTRACTDirect learning control (DLC) schemes have been developed recently to address non‐repeatable trajectory tracking problems. Unlike conventional iterative learning schemes, DLC schemes learn a set of unknown basis function vectors which can be used to generate the desired control profile of a new trajectory. DLC schemes use all available trajectory tracking information to obtain the unknown basis function vectors in a Least Squares and pointwise manner. A drawback of DLC is that the inverse matrix calculation is inevitable, which is time consuming and may result in singularities due to the batch processing nature. A Recursive Direct Learning Control method is proposed which learns the basis function vectors meanwhile overcomes the implementation difficulties in DLC schemes. The focus of this paper is on learning the control profile of trajectories with same operation period but different magnitude scales. The recursive learning method makes use of one trajectory information at a time, thus avoids the batch processing. The scheme is first developed for a class of nonlinear time varying systems and then extended to cover more general classes of nonlinear systems including robotic manipulator dynamics. Extensive simulation results on a two‐link robotic model are provided to confirm the features of the proposed algorithm.

Direct Model Reference Learning and Repetitive Control

Abstract Iterative learning and repetitive controllers learn to improve their performance when repeatedly executing a task. The learning process can be based on many different convergence concepts. However, the similarity of objective with adaptive control suggest the use of adaptive concepts. Here direct model reference adaptive control ideas are converted to produce discrete-time direct model reference learning control and direct model reference repetitive control laws with guaranteed convergence to zero tracking error. These control laws are used to modify the command going into a feedback control system, converging on whatever command is necessary to produce the desired output. The original adaptive control concepts used are for linear time invariant systems without deterministic disturbances, but the learning and repetitive control laws developed handle time varying systems with repetitive disturbances. Such general models are important since they can represent the behavior of nonlinear systems linea...

EXTENDED DIRECT LEARNING CONTROL FOR MULTI-INPUT MULTI-OUTPUT NONLINEAR SYSTEMS

For a class of nonlinear systems which perform a given task repetitively, an extended type of a direct learning control (DLC) is proposed using the information on the (vector) relative degree of the a multi-input multi-output system. DLC method can generate the desired control input directly from prestored control input profiles with different time scales without any repetitive learning process. It is shown that existing DLC methods can be applied only to a certain limited class of nonlinear systems and the information on the relative degree of a nonlinear system is essential to find the desired control input if the system has higher relative degree.

Extended direct learning control of systemswith arbitrary relative degree

An extended direct learning control scheme is proposed for systems the transfer functions of which have an arbitrary relative degree. The direct learning control scheme generates the desired control input, which yields the given desired output trajectory, by using prestored control input profiles without the need for a repetitive learning process. It is shown that the information on the relative degree of a system can be used to extend the class of systems to which direct learning control is applicable.

Multi-scale direct learning control of linear time-varying high-order systems

In this paper, a multi-scale direct learning control (DLC) scheme is proposed. DLC is defined as the direct generation of the desired control profile from existing control inputs without any repeated learning. DLC schemes developed hitherto only consider trajectories with a single-scale relation in that any two trajectories are spatially proportional to each other with a single scale for all coordinates (dimensions). In comparison, the new scheme is capable of handling trajectories with multiple scales, i.e., each spatial coordinate of a trajectory is proportional to that of another trajectory associated with a distinct scale. It is shown that the new scheme is applicable to linear time-varying high-order systems. Simulation results demonstrate its effectiveness.

Multiple magnitude scale DLC scheme for high order systems

Abstract In this paper, we propose two multi-scale Direct Learning Control (DLC) schemes for time-varying high order systems, one for linear systems and the other for nonlinear systems. DLC schemes developed hitherto only consider trajectories with a single-scale relation in that any two trajectories are spatially proportional to each other with a single scale for all coordinates (dimensions). In comparison, the new schemes are capable of handling trajectories with multiple scales, i.e., each spatial coordinate of a trajectory is proportional to that of another trajectory associated with a distinct scale. Simulation results confirm the effectiveness of these methods.

On the integrated learning control method

A new integrated learning control scheme is developed by integration of iterative learning, adaptive robust and direct learning control. By factorizing the system uncertainties into known scaling matrix and unknown basis function vectors as in direct learning control, the new learning control scheme can deal with non-uniform trajectory tracking problems. By incorporating the adaptive robust control scheme, the integrated learning system is able to handle both structured and unstructured system uncertainties. By means of Lyapunov's direct method, the integrated learning control is analyzed for a class of non-linear and time varying systems.

Dual-scale direct learning control of trajectory tracking for a class of nonlinear uncertain systems

In this paper, a new learning control scheme: dual-scale direct learning control for trajectory tracking problems is presented. Direct learning control (DLC) is defined as the direct generation of the desired control input profile from existing control input profiles without any repetitive learning. For a class of nonlinear uncertain dynamic systems, the proposed dual-scale direct learning control scheme is able to generate the desired control input profile for a new trajectory which differs from all the previously controlled and stored trajectories in both time and magnitude scales.

Dual-scale direct learning control of trajectory tracking for a class of nonlinear uncertain systems

In this paper the authors present a new learning control scheme: dual-scale direct learning control for trajectory tracking problems. Direct learning control (DLC) is defined as the direct generation of the desired control input profile from existing control input profiles without any repetitive learning. The DLC schemes developed hitherto are capable of learning the desired control signals for the new trajectory from past tracking control profiles which correspond to trajectories with either nonuniform time scales or nonuniform magnitude scales. Here the authors further show that for a class of nonlinear uncertain dynamic systems, the proposed dual-scale DLC scheme is able to generate the desired control input profile for a new trajectory which differs from all of the previously controlled and stored trajectories in both time and magnitude scales.

Direct learning of control efforts for trajectories with different time scales

We introduce a new learning control method, direct learning control, which is defined as the generation of the desired control input profile directly from existing control input profiles without any repeated learning. The motivation of developing direct learning control schemes is to overcome the limitation of conventional learning control methods which require that the desired tracking patterns (trajectories) be strictly identical (repeatable) throughout the learning process. There are two main advantages of the direct learning control method. The first is that the learning control system is capable of fully utilizing the prestored control input signals which may correspond to tracking patterns with different time scales and be achieved through various control approaches. The second is the direct generation of the desired control input profile; thereafter it is possible to remove the whole iterative learning process. The focus of this paper is on direct learning of a class of nonperiodic trajectories which are identical in spatial distribution but different in time scales.

Direct learning of control efforts for trajectories with different magnitude scales

In this paper, we propose a new concept— direct learning which is defined as the generation of the desired control input profile directly from existing control input profiles without any repeated learning. The motivation of developing direct learning control schemes is to overcome the limitation of conventional learning control methods which require that the desired tracking patterns (trajectories) be strictly repeatable throughout the learning process. The main advantages of the direct learning are (1) the capability of fully utilizing the preobtained control input signals which may correspond to tracking patterns with different magnitude scales and be achieved through various control approaches; (2) direct generation of the desired control input profile without repeating the operation cycles. The focus of this paper is on direct learning for a class of trajectories which have identical operation periods but are different in magnitude scales.