Bettering operation of Robots by learning

Arimoto, Suguru; Kawamura, Sadao; Miyazaki, Fumio

doi:10.1002/rob.4620010203

Cited by 3,060 publications

(1,407 citation statements)

References 4 publications

Supporting

Mentioning

1,397

Contrasting

Unclassified

Order By: Relevance

“…In this case the real-time control system is typically some type of asymptotic tracking control system, and ILC then uses the results of the repeated tracking exercises to further improve the tracking accuracy (for early references see, e.g., [2] and [3]). This is done on an ongoing basis, and therefore ILC not only optimizes tracking accuracy for a given desired output and one specific set of system dynamics, but can also be used to maintain tracking accuracy by responding to gradual changes in the system dynamics through its "learning" capability.…”

Section: Introductionmentioning

confidence: 99%

An alternative update formula for non-linear model-based iterative learning control

Eksteen

Heyns

2015

Inverse Problems in Science and Engineering

View full text Add to dashboard Cite

The conventional Iterative Learning Control (ILC) algorithm for model-based ILC of nonlinear systems is presented with use of a nonlinear inverse model as ILC compensator. The nonlinear inverse model is solved with stable inversion. In addition an alternative ILC algorithm for model-based ILC of nonlinear systems is developed, also with using a nonlinear inverse model as ILC compensator. Some connections between the conventional and alternative ILC algorithms and Picard, Mann and Ishikawa iteration are explored. In a number of theoretical examples the conventional and alternative algorithms are compared.

show abstract

Section: Introductionmentioning

confidence: 99%

An alternative update formula for non-linear model-based iterative learning control

Eksteen

Heyns

2015

Inverse Problems in Science and Engineering

View full text Add to dashboard Cite

show abstract

“…ILC was first proposed in the robotics community as a means to successively reduce the tracking errors in repetitive dynamic processes (Arimoto et al, 1984). In the ILC literature, the run-to-run aspect was developed before studying its interaction with the on-line (withinrun) adaptation.…”

Section: Trajectory Tracking Using Ilcmentioning

confidence: 99%

Combined On-Line and Run-to-Run Optimization of Batch Processes with Terminal Constraints

Welz

Srinivasan

Bonvin

2004

IFAC Proceedings Volumes

View full text Add to dashboard Cite

This paper describes the optimization of batch processes in the presence of uncertainty and constraints. The optimal solution consists of keeping certain path and terminal constraints active and driving the sensitivities to zero. The case where the terminal constraints have a larger bearing on the cost than the sensitivities is considered, for which a two-time-scale methodology is proposed. The problem of meeting the active terminal constraints is addressed on-line using trajectory tracking, whilst pushing the sensitivities to zero is implemented on a run-to-run basis. The paper also discusses the run-to-run improvement of trajectory tracking via iterative learning control. The proposed methodology is illustrated in simulation on a batch distillation system.

show abstract

“…Iterative learning control (ILC) algorithm is a kind of non-model-dependent control method which is suitable for the object with repetitive motion characteristics. It was proposed by Arimoto et al in 1984 and has been applied to various fields of control theory and engineering recently. It focuses on the problems where the interaction between different durations is normally zero but the repetition of tracking the same trajectory creates a possibility of improving tracking performance [3].…”

Section: Introductionmentioning

confidence: 99%

“…When ILC was first proposed, only differential item of the error was used to correct the last input, i.e. D-type ILC [4]. After that, some other scholars take proportional and integral items into account using PID control for reference, and propose ILC methods of PI-type, PD-type and PID-type.…”

Section: Introductionmentioning

confidence: 99%

Fuzzy PD-Type Iterative Learning Control of a Single Pneumatic Muscle Actuator

Meng

et al. 2017

Intelligent Robotics and Applications

View full text Add to dashboard Cite

Abstract. Pneumatic muscles actuator (PMA) is widely used in the field of rehabilitation robot for its good flexibility, light weight and high power/mass ratio as compared to traditional actuator. In this paper, a fuzzy logic-based PD-type iterative learning controller (ILC) is proposed to control the PMA to track a predefined trajectory more precisely during repetitive movements. In order to optimize the parameters of the learning law, fuzzy logic control is introduced into ILC to achieve smaller errors and faster convergence. A simulation experiment was first conducted by taking the PMA model fitted by support vector machine (SVM) as controlled target, which showed that the proposed method achieved a better tracking performance than traditional PD-type ILC. A satisfactory control effect was also obtained when fuzzy PD-type ILC was applied to actual PMA control experiment. Result showed that it takes 25 iterations for the maximum error of trajectory converges to a minimum of about 0.2.

show abstract

Bettering operation of Robots by learning

Cited by 3,060 publications

References 4 publications

An alternative update formula for non-linear model-based iterative learning control

An alternative update formula for non-linear model-based iterative learning control

Combined On-Line and Run-to-Run Optimization of Batch Processes with Terminal Constraints

Fuzzy PD-Type Iterative Learning Control of a Single Pneumatic Muscle Actuator

Contact Info

Product

Resources

About