Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4792864
G. Puskorius, L. Feldkamp
This paper describes the development of recurrent neural network controllers for an automotive engine idle speed control (ISC) problem. Engine ISC is a difficult problem because of troublesome characteristics such as severe process nonlinearities, variable time delays, time-varying process dynamics and unobservable system states and disturbances. We demonstrate that recurrent neural network controllers can be trained to handle these difficulties gracefully while achieving good regulator performance for a representative model of 4-cylinder, 1.6 liter engine. Empirical results clearly illustrate that neural network controllers with relatively large amounts of internal feedback provide more robust performance for the ISC problem than do neural network controllers that are static or contain limited internal recurrent connections.
{"title":"Automotive Engine Idle Speed Control with Recurrent Neural Networks","authors":"G. Puskorius, L. Feldkamp","doi":"10.23919/ACC.1993.4792864","DOIUrl":"https://doi.org/10.23919/ACC.1993.4792864","url":null,"abstract":"This paper describes the development of recurrent neural network controllers for an automotive engine idle speed control (ISC) problem. Engine ISC is a difficult problem because of troublesome characteristics such as severe process nonlinearities, variable time delays, time-varying process dynamics and unobservable system states and disturbances. We demonstrate that recurrent neural network controllers can be trained to handle these difficulties gracefully while achieving good regulator performance for a representative model of 4-cylinder, 1.6 liter engine. Empirical results clearly illustrate that neural network controllers with relatively large amounts of internal feedback provide more robust performance for the ISC problem than do neural network controllers that are static or contain limited internal recurrent connections.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"12 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134333182","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4792871
P. Hsu
The main problem of controlling a multiple manipulator system is the conflicting actions between the manipulators. This conflicting action is often due to geometric modeling error or, in the case of independently controlled system, mismatch between the reference trajectories of the controllers. The proposed control scheme resolves the conflicting actions via a `mutual learning' process. The adaptive mechanism of each controller modifies its reference trajectory so as to adapt to others. It is shown that a common reference trajectory will eventually be reached by all the controllers. This adaptation process is carried out without explicit communication between the controllers. The proposed scheme was verified by computer simulations and experiments on a dual-manipulator system.
{"title":"Adaptive Resolution of Conflicts in Multimanipulator Systems","authors":"P. Hsu","doi":"10.23919/ACC.1993.4792871","DOIUrl":"https://doi.org/10.23919/ACC.1993.4792871","url":null,"abstract":"The main problem of controlling a multiple manipulator system is the conflicting actions between the manipulators. This conflicting action is often due to geometric modeling error or, in the case of independently controlled system, mismatch between the reference trajectories of the controllers. The proposed control scheme resolves the conflicting actions via a `mutual learning' process. The adaptive mechanism of each controller modifies its reference trajectory so as to adapt to others. It is shown that a common reference trajectory will eventually be reached by all the controllers. This adaptation process is carried out without explicit communication between the controllers. The proposed scheme was verified by computer simulations and experiments on a dual-manipulator system.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131771582","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4792836
B. R. Patnaik, G. Heppler, D. Wang
Piezoelectric actuators are capable of generating a distributed moment, and can be used to control the motion of a flexible structure. A simple piesoelectric damper is considered here for controlling the vibration of a cantilever Euler-Bernoulli beam. Asymptotic stability of the closed loop system is established via a distributed parameter extension of Liapunov's direct method.
{"title":"Stability Analysis of a Piezoelectric Vibration Controller for an Euler-Bernoulli Beam","authors":"B. R. Patnaik, G. Heppler, D. Wang","doi":"10.23919/ACC.1993.4792836","DOIUrl":"https://doi.org/10.23919/ACC.1993.4792836","url":null,"abstract":"Piezoelectric actuators are capable of generating a distributed moment, and can be used to control the motion of a flexible structure. A simple piesoelectric damper is considered here for controlling the vibration of a cantilever Euler-Bernoulli beam. Asymptotic stability of the closed loop system is established via a distributed parameter extension of Liapunov's direct method.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131820471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4792894
G. Shi, Xiang Liu
The problem of circular pole assignment for discrete-time singular systems is considered. The goal of the problem is to assign the maximum number of finite eigenvalues in a prespecified circle and guarantee the closed-loop regularity. A simple, effective generalized Riccati equation approach is developed to solve the addressed problem. It is shown that a desired state feedback law is determined by using the solution of a standard discrete Riccati equation which can be computed directly.
{"title":"Circular Pole Assignment for Discrete-Time Singular Systems","authors":"G. Shi, Xiang Liu","doi":"10.23919/ACC.1993.4792894","DOIUrl":"https://doi.org/10.23919/ACC.1993.4792894","url":null,"abstract":"The problem of circular pole assignment for discrete-time singular systems is considered. The goal of the problem is to assign the maximum number of finite eigenvalues in a prespecified circle and guarantee the closed-loop regularity. A simple, effective generalized Riccati equation approach is developed to solve the addressed problem. It is shown that a desired state feedback law is determined by using the solution of a standard discrete Riccati equation which can be computed directly.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130751589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.1109/ACC.1993.4176134
P. Khargonekar, M. Rotea, N. Sivashankar
In this paper, we consider a multiple objective control problem. If the exogenous input matrices in a state-space model of the plant under control satisfy a generic rank condition, we show that all the individual state-feedback controllers which achieve desirable performance and robustness levels (as measured by suitable closed loop transfer matrices) can be combined to generate a single state-feedback controller that simultaneously achieves the same performance and robustness levels. In the output feed-back case we show how to recover (to any degree of accuracy) all the state-feedback closed loop properties with a single observer based controller when the subsystem from the exogenous input to the measured output satisfies a minimum phase assumption.
{"title":"Exact and Approximate Solutions to a Class of Multiobjective Controller Synthesis Problems","authors":"P. Khargonekar, M. Rotea, N. Sivashankar","doi":"10.1109/ACC.1993.4176134","DOIUrl":"https://doi.org/10.1109/ACC.1993.4176134","url":null,"abstract":"In this paper, we consider a multiple objective control problem. If the exogenous input matrices in a state-space model of the plant under control satisfy a generic rank condition, we show that all the individual state-feedback controllers which achieve desirable performance and robustness levels (as measured by suitable closed loop transfer matrices) can be combined to generate a single state-feedback controller that simultaneously achieves the same performance and robustness levels. In the output feed-back case we show how to recover (to any degree of accuracy) all the state-feedback closed loop properties with a single observer based controller when the subsystem from the exogenous input to the measured output satisfies a minimum phase assumption.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131111395","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4792917
M. Kabuli, S. Bhat, R. Kosut
A real-time control problem associated with a flexible testbed fixture is considered. The regulated variable is the deflection at the tip of a flexible beam attached to an inertia wheel base which is subject to a disturbance torque. The control torque affects the base wheel through a coupling which has both compliance and variable backlash. The performance specifications include disturbance attenuation and rapid slewing at the tip. The designs are based on an analytically derived continuous-time model which is tuned in accordance with the measured data. The model consists of an interconnection of a linear time-invariant part and piecewise-linear algebraic nonlinearities modeling friction and backlash. The candidate designs must be based on available measurements only; more-over, the final discrete-time control law must be executable with the real-time controller hardware limitations at hand.
{"title":"Real-Time Implementation Issues in Nonlinear Model Inversion","authors":"M. Kabuli, S. Bhat, R. Kosut","doi":"10.23919/ACC.1993.4792917","DOIUrl":"https://doi.org/10.23919/ACC.1993.4792917","url":null,"abstract":"A real-time control problem associated with a flexible testbed fixture is considered. The regulated variable is the deflection at the tip of a flexible beam attached to an inertia wheel base which is subject to a disturbance torque. The control torque affects the base wheel through a coupling which has both compliance and variable backlash. The performance specifications include disturbance attenuation and rapid slewing at the tip. The designs are based on an analytically derived continuous-time model which is tuned in accordance with the measured data. The model consists of an interconnection of a linear time-invariant part and piecewise-linear algebraic nonlinearities modeling friction and backlash. The candidate designs must be based on available measurements only; more-over, the final discrete-time control law must be executable with the real-time controller hardware limitations at hand.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133508050","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4793487
S. Phillips, Kuo-Chen Chou
In this paper constant gain feedforward is used to minimize the transient tracking error for the linear quadratic based robust multivariable servomechanism. This approach augments the internal model control and quadratic cost minimization with feedforward gains which minimize the infinite time horizon tracking error.
{"title":"Feedforward Augmentation of Internal Model Control for Tracking","authors":"S. Phillips, Kuo-Chen Chou","doi":"10.23919/ACC.1993.4793487","DOIUrl":"https://doi.org/10.23919/ACC.1993.4793487","url":null,"abstract":"In this paper constant gain feedforward is used to minimize the transient tracking error for the linear quadratic based robust multivariable servomechanism. This approach augments the internal model control and quadratic cost minimization with feedforward gains which minimize the infinite time horizon tracking error.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133145576","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4793322
C. Blackwell
Lyapunov's Direct Principle is applied to the problem of establishing bounds of convergence which result from the saturation of actuators. The treatment is applicable to those cases for which a linear time invariant nominal model is appropriate. The results are demonstrated with a four-state example.
{"title":"Actuator Saturation and Control System Performance","authors":"C. Blackwell","doi":"10.23919/ACC.1993.4793322","DOIUrl":"https://doi.org/10.23919/ACC.1993.4793322","url":null,"abstract":"Lyapunov's Direct Principle is applied to the problem of establishing bounds of convergence which result from the saturation of actuators. The treatment is applicable to those cases for which a linear time invariant nominal model is appropriate. The results are demonstrated with a four-state example.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"38 21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131697307","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4793196
H. Berenji, P. Khedkar
Non-adaptive fuzzy logic controllers can become adaptive by learning from experience in the framework of reinforcement learning. In this paper, we discuss fuzzy reinforcement learning as a hybrid approach which provides a unified framework for including two types of prior knowledge: knowledge for control action selection and knowledge for performance evaluation. We describe GARIC, an architecture for combining fuzzy logic control and reinforcement learning, and apply it to cart-pole balancing and the Space Shuttle attitude control.
{"title":"Adaptive Fuzzy Control with Reinforcement Learning","authors":"H. Berenji, P. Khedkar","doi":"10.23919/ACC.1993.4793196","DOIUrl":"https://doi.org/10.23919/ACC.1993.4793196","url":null,"abstract":"Non-adaptive fuzzy logic controllers can become adaptive by learning from experience in the framework of reinforcement learning. In this paper, we discuss fuzzy reinforcement learning as a hybrid approach which provides a unified framework for including two types of prior knowledge: knowledge for control action selection and knowledge for performance evaluation. We describe GARIC, an architecture for combining fuzzy logic control and reinforcement learning, and apply it to cart-pole balancing and the Space Shuttle attitude control.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131858853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1993-06-02DOI: 10.23919/ACC.1993.4793427
S. Srinivasan, K. Moore, D. Naidu
In this paper we present some preliminary ideas for the design of a continuous nonlinear neural networks with "learning." Specifically, we introduce the idea of learning in Hopfield recursive neural networks. The network is trained so that application of a set of inputs produces the desired set of outputs. A method is developed to determine the interconnecting weights for the network, so as to achieve the desired stable equilibrium points. Also, this method illustrates a way to 'learn' the interconnecting weights that are not computed a priori. Conditions are obtained for the asymptotic stability of the equilibrium points. An illustrative simulation is presented.
{"title":"An Approach to Learning in Hopfield Neural Networks","authors":"S. Srinivasan, K. Moore, D. Naidu","doi":"10.23919/ACC.1993.4793427","DOIUrl":"https://doi.org/10.23919/ACC.1993.4793427","url":null,"abstract":"In this paper we present some preliminary ideas for the design of a continuous nonlinear neural networks with \"learning.\" Specifically, we introduce the idea of learning in Hopfield recursive neural networks. The network is trained so that application of a set of inputs produces the desired set of outputs. A method is developed to determine the interconnecting weights for the network, so as to achieve the desired stable equilibrium points. Also, this method illustrates a way to 'learn' the interconnecting weights that are not computed a priori. Conditions are obtained for the asymptotic stability of the equilibrium points. An illustrative simulation is presented.","PeriodicalId":162700,"journal":{"name":"1993 American Control Conference","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131863606","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：