동적 평면 제어기법을 이용한 탄-소성 마찰모델 기반의 비선형 서보시스템 정밀위치제어 = Precise position control of nonlinear servo system using dynamic surface control method based on elasto-plastic friction model

다국어 초록 (Multilingual Abstract)

This paper presents precise position control method of nonlinear servo system with nonlinear friction and uncertainty and the performance of proposed position control system was evaluated from simulation and experiment.
One of the problem to overcome to implement high precision servo control is friction which exists between the contact surfaces of two materials. Friction phenomenon has been studied by many researchers and many friction model in order to capture the unique friction properties have been improved or developed. LuGre friction model is representative friction model and many friction compensation schemes based on LuGre friction have been presented in order to alleviate the tracking performance.
However, LuGre has shortcoming that do not exactly presents stiction and pre-sliding displacement which happen in pre-sliding region. Elasto-plastic friction model was developed in order to overcome this problem. Elasto-plastic model displays friction phenomenon by dividing to elastic deformation section and plastic deformation section. In the Elasto-plastic model, interior state variable and pre-sliding displacement are important design parameters which decide friction properties. In order to implement friction compensation scheme based on mathematical friction model, identification of parameters for friction model is required. Identification of friction parameters is very difficult to require much times and effort. In addition, even with successful completion of the friction identification process, it is difficult to achieve desirable tracking performance due to time-varying characteristics of the nonlinear friction which depend on precision manufacturing, lubrication, temperature, and contamination. Therefore, friction parameter observation technique is needed to apply friction model in precision position control system. Although precision control schemes are very effective for compensating nonlinear friction, the performance of position control system can be limited by system uncertainty which is occurred by some assumption or neglected disturbance in modeling process. To deal with unknown system uncertainties effectively, artificial intelligent(AI) schemes such as fuzzy and neural networks algorithms can be used to compensate system uncertainties. AI schemes can effectively approximate any continuous function. Application of estimation methods for nonlinear friction and system uncertainties should be achieved carefully. Over-estimation for unknown parameters or uncertainties can influence in stability of position control system. Therefore, dynamic surface control method is adopt to consider over estimation problem of friction observer and system uncertainty estimator at position controller design process. Traditional back-stepping algorithms, although systematic, suffer from an ‘explosion of complexity’ due to the necessity to perform repeated differentiations of the nonlinear functions. Because dynamic surface control scheme uses first-order filtering of the synthetic inputs at each level of the traditional back-stepping approach, organization of control input is simple than control input of back-stepping control scheme.
In order to evaluate proposed position control system, simulation and experiment carry out for scorbot robot system and ball-screw system.

목차 (Table of Contents)

• 1 장 서론 1
• 1.1 연구배경 1
• 1.2 연구동향 3
• 1.3 연구목적 및 내용 6
• 1.4 논문의 구성 8
• 2 장 비선형 마찰 모델링 9
• 2.1 개요 9
• 2.2 비선형 마찰의 발생 메카니즘 10
• 2.3 마찰 모델 및 특성 13
• 2.4 마찰 모델 비교 22
• 2.5 요약 27
• 3 장 마찰 보상 위치제어시스템 설계 28
• 3.1 개요 28
• 3.2 마찰 관측기 설계 29
• 3.3 퍼지-신경망 설계 32
• 3.4 적응백스테핑 제어기 설계 37
• 3.5 동적 평면 제어기 설계 47
• 3.6 요약 58
• 4 장 위치제어 성능 평가 60
• 4.1 개요 60
• 4.2 Scorbot 로봇시스템의 위치제어 61
• 4.3 볼-스크류 시스템의 위치제어 85
• 4.4 결론 및 고찰 94
• 5 장 결론 및 향후 과제 95
• 5.1 결론 95
• 5.2 향후 과제 96
• REFERENCE 97
• ABSTRACT 102