Pub Date : 2011-09-29DOI: 10.1109/MMAR.2011.6031338
Urszula Libal
For high-dimensional data processing, like pattern recognition, it seems desirable to precede with a reduction of the number of describing features. Our aim is a comparison of various feature selection methods for pattern recognition. We consider two-class supervised classification problem for signals decomposed in wavelet bases. We test kNN classification rule with soft and hard thresholding, performed in two stages: (1) wavelet detail coefficient thresholding (noise reduction) and (2) searching for the most differentiating coefficients between classes (selection of discriminating coefficients). We present a new classification rule based on LARS/LASSO. We compare criteria for L1-norm regularization of wavelet coefficients: AIC, BIC and the thresh derived for kNN rule. There were performed simulations for noisy signals with SNR in the range from 0 to 22 [dB], approximated for all possible wavelet resolutions. The quality of pattern recognition for the presented algorithms was measured by the estimated recognition risk and the size of reduced model.
{"title":"Feature selection for pattern recognition by LASSO and thresholding methods - a comparison","authors":"Urszula Libal","doi":"10.1109/MMAR.2011.6031338","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031338","url":null,"abstract":"For high-dimensional data processing, like pattern recognition, it seems desirable to precede with a reduction of the number of describing features. Our aim is a comparison of various feature selection methods for pattern recognition. We consider two-class supervised classification problem for signals decomposed in wavelet bases. We test kNN classification rule with soft and hard thresholding, performed in two stages: (1) wavelet detail coefficient thresholding (noise reduction) and (2) searching for the most differentiating coefficients between classes (selection of discriminating coefficients). We present a new classification rule based on LARS/LASSO. We compare criteria for L1-norm regularization of wavelet coefficients: AIC, BIC and the thresh derived for kNN rule. There were performed simulations for noisy signals with SNR in the range from 0 to 22 [dB], approximated for all possible wavelet resolutions. The quality of pattern recognition for the presented algorithms was measured by the estimated recognition risk and the size of reduced model.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115765985","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031324
W. Hunek, R. Stanisławski, K. Latawiec, M. Lukaniszyn
This paper presents (structurally stable) pole-free control designs for (a deterministic version of) generalized minimum variance control and linear quadratic regulation for nonsquare LTI MIMO systems. For both discrete-time control strategies, two alternative approaches are offered, the numerical one related with limiting control solutions as control weighting matrices tend to the zero ones and the other produced by analytically-derived perfect controls, for which it is possible to put the weighting matrices directly to the zero ones. In the latter case, the solutions are identical with that for minimum variance control, in which case the pole-free designs must account for the so-called control zeros, being the extension of transmission zeros for nonsquare LTI MIMO systems. Simulation examples illustrate the theoretical achievements of the paper.
{"title":"Pole-free GMVC and LQR designs for nonsquare LTI MIMO systems","authors":"W. Hunek, R. Stanisławski, K. Latawiec, M. Lukaniszyn","doi":"10.1109/MMAR.2011.6031324","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031324","url":null,"abstract":"This paper presents (structurally stable) pole-free control designs for (a deterministic version of) generalized minimum variance control and linear quadratic regulation for nonsquare LTI MIMO systems. For both discrete-time control strategies, two alternative approaches are offered, the numerical one related with limiting control solutions as control weighting matrices tend to the zero ones and the other produced by analytically-derived perfect controls, for which it is possible to put the weighting matrices directly to the zero ones. In the latter case, the solutions are identical with that for minimum variance control, in which case the pole-free designs must account for the so-called control zeros, being the extension of transmission zeros for nonsquare LTI MIMO systems. Simulation examples illustrate the theoretical achievements of the paper.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127329191","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031350
G. Barna
The subject of this paper is a theoretical analysis of Wheel Slide Protection (WSP) controllers for rail vehicles. First of all a principle of operation of WSP systems has been described and problems concerning designing WSP controllers have been specified. Then structure of a WSP controller compared to a classic control diagram is presented. Various types of WSP controllers are compared and their advantages and disadvantages are presented. Conclusions concerning adopting structure and parameters of controllers are drawn. Further investigations aimed at facilitating development of efficient WSP controllers are proposed.
{"title":"Theoretical analysis of Wheel Slide Protection controllers for rail vehicles","authors":"G. Barna","doi":"10.1109/MMAR.2011.6031350","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031350","url":null,"abstract":"The subject of this paper is a theoretical analysis of Wheel Slide Protection (WSP) controllers for rail vehicles. First of all a principle of operation of WSP systems has been described and problems concerning designing WSP controllers have been specified. Then structure of a WSP controller compared to a classic control diagram is presented. Various types of WSP controllers are compared and their advantages and disadvantages are presented. Conclusions concerning adopting structure and parameters of controllers are drawn. Further investigations aimed at facilitating development of efficient WSP controllers are proposed.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116540238","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031327
S. Nikiel, Pawel Dabrowski
In this paper, a novel information design algorithm is proposed. It is based on the simulated annealing, which is regarded as one of the important meta-heuristic optimization methods. The idea behind our algorithm originates from the search for optimal solution, where selection avoids local minima. The simulated annealing model is used in the construction of diagrams depicting heating systems. The method offers an automated visual design tool, useful in CAD1 construction of complex systems. The paper discusses the theoretical background, the properties of the proposed algorithm alongside with the sample application prototype outputs.
{"title":"Deployment algorithm using simulated annealing","authors":"S. Nikiel, Pawel Dabrowski","doi":"10.1109/MMAR.2011.6031327","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031327","url":null,"abstract":"In this paper, a novel information design algorithm is proposed. It is based on the simulated annealing, which is regarded as one of the important meta-heuristic optimization methods. The idea behind our algorithm originates from the search for optimal solution, where selection avoids local minima. The simulated annealing model is used in the construction of diagrams depicting heating systems. The method offers an automated visual design tool, useful in CAD1 construction of complex systems. The paper discusses the theoretical background, the properties of the proposed algorithm alongside with the sample application prototype outputs.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115387370","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031326
T. Duncan, B. Maslowski, B. Pasik-Duncan
A control problem for a linear system in a Hilbert space with a fractional Brownian motion and a quadratic cost in the state and the control is solved. The feedback form of the optimal control and the optimal cost are given. The optimal control is the sum of the well known linear feedback control for the associated deterministic linear-quadratic control problem and a suitable prediction of an optimal system response to the future noise. Some examples of controlled stochastic partial differential equations are given.
{"title":"Control of some linear stochastic systems in a Hilbert space with fractional Brownian motions","authors":"T. Duncan, B. Maslowski, B. Pasik-Duncan","doi":"10.1109/MMAR.2011.6031326","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031326","url":null,"abstract":"A control problem for a linear system in a Hilbert space with a fractional Brownian motion and a quadratic cost in the state and the control is solved. The feedback form of the optimal control and the optimal cost are given. The optimal control is the sum of the well known linear feedback control for the associated deterministic linear-quadratic control problem and a suitable prediction of an optimal system response to the future noise. Some examples of controlled stochastic partial differential equations are given.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128975810","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031353
A. Rauh, Luise Senkel, Christina Dittrich, H. Aschemann
Model-based control strategies for pressure boosting systems in water supply for high-rise buildings require the knowledge of internal system states and parameters which are not directly accessible for measurements with reasonable effort in practical applications. For that reason, the observer design is investigated in this paper for finite-dimensional system models of the dynamics of water supply systems. The corresponding state equations are derived as generalized network models which represent the overall system dynamics. The generalized network model involves pump-motor units as power sources, hydraulic capacitances, and fluid inertia which are described by separate sets of ordinary differential equations. These sets of ordinary differential equations are coupled by algebraic constraints derived from suitable balance equations. The applicability of a model-based state observer for the reconstruction of the volume flow rates in the individual pipeline segments as well as the pressure distribution along the fluidic network is demonstrated by numeric simulations. Moreover, the identification of nonlinear volume flow-dependent characteristics for fluidic resistances is demonstrated on the basis of an online applicable disturbance observer. Both observer models contain basic assumptions on the fluidic resistances which are refined online with the help of the presented techniques.
{"title":"Observer design for state and parameter estimation in pressure boosting systems for water supply","authors":"A. Rauh, Luise Senkel, Christina Dittrich, H. Aschemann","doi":"10.1109/MMAR.2011.6031353","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031353","url":null,"abstract":"Model-based control strategies for pressure boosting systems in water supply for high-rise buildings require the knowledge of internal system states and parameters which are not directly accessible for measurements with reasonable effort in practical applications. For that reason, the observer design is investigated in this paper for finite-dimensional system models of the dynamics of water supply systems. The corresponding state equations are derived as generalized network models which represent the overall system dynamics. The generalized network model involves pump-motor units as power sources, hydraulic capacitances, and fluid inertia which are described by separate sets of ordinary differential equations. These sets of ordinary differential equations are coupled by algebraic constraints derived from suitable balance equations. The applicability of a model-based state observer for the reconstruction of the volume flow rates in the individual pipeline segments as well as the pressure distribution along the fluidic network is demonstrated by numeric simulations. Moreover, the identification of nonlinear volume flow-dependent characteristics for fluidic resistances is demonstrated on the basis of an online applicable disturbance observer. Both observer models contain basic assumptions on the fluidic resistances which are refined online with the help of the presented techniques.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127589683","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031365
R. Beckmann, W. Drewelow
A physically based, nonlinear model of the masses of air and burned gas inside the cylinders of a spark ignition (SI) engine is presented. Goal of this work is to develop a feed forward control law, which perfectly tracks given reference values of these masses. This is done by utilizing the exact linearization approach [1]. The final feed forward control law is verified in combination with a slightly modified design model as well as the original model.
{"title":"Modeling and exact feed forward control of air and burned gas fraction within SI engine cylinders","authors":"R. Beckmann, W. Drewelow","doi":"10.1109/MMAR.2011.6031365","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031365","url":null,"abstract":"A physically based, nonlinear model of the masses of air and burned gas inside the cylinders of a spark ignition (SI) engine is presented. Goal of this work is to develop a feed forward control law, which perfectly tracks given reference values of these masses. This is done by utilizing the exact linearization approach [1]. The final feed forward control law is verified in combination with a slightly modified design model as well as the original model.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128219260","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031312
Z. Emirsajlow, A. Krakowiak, A. Kowalewski, J. Sokołowski
In the paper the first order sensitivity analysis is performed for a class of optimal control problems for parabolic-hyperbolic equations. A singular perturbation of geometrical domain of integration is introduced in the form of a circular hole. The Steklov-Poincaré operator on a circle is defined in order to reduce the problem to regular perturbations in the truncated domain. The optimality system is differentiated with respect to the small parameter and the directional derivative of the optimal control is obtained as a solution to an auxiliary optimal control problem.
{"title":"Sensitivity analysis of parabolic-hyperbolic optimal control problems","authors":"Z. Emirsajlow, A. Krakowiak, A. Kowalewski, J. Sokołowski","doi":"10.1109/MMAR.2011.6031312","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031312","url":null,"abstract":"In the paper the first order sensitivity analysis is performed for a class of optimal control problems for parabolic-hyperbolic equations. A singular perturbation of geometrical domain of integration is introduced in the form of a circular hole. The Steklov-Poincaré operator on a circle is defined in order to reduce the problem to regular perturbations in the truncated domain. The optimality system is differentiated with respect to the small parameter and the directional derivative of the optimal control is obtained as a solution to an auxiliary optimal control problem.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133715295","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031374
B. Kreczmer
In this paper an approach to gesture recognition is presented. It is based on application of simple ultrasonic range-finders. A set of five gestures is described which can be distinguish in this way. The described method has been implemented for a mobile robot PeopleBot.
{"title":"Gestures recognition by using ultrasonic range-finders","authors":"B. Kreczmer","doi":"10.1109/MMAR.2011.6031374","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031374","url":null,"abstract":"In this paper an approach to gesture recognition is presented. It is based on application of simple ultrasonic range-finders. A set of five gestures is described which can be distinguish in this way. The described method has been implemented for a mobile robot PeopleBot.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123970390","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 : 2011-09-29DOI: 10.1109/MMAR.2011.6031322
A. Kowalewski
Various optimization problems for linear parabolic systems with multiple constant time delays are considered. In this paper, we consider an optimal distributed control problem for a linear infinite order parabolic system in which multiple constant time delays appear in the state equation. Sufficient conditions for the existence of a unique solution of the infinite order parabolic time delay equation with the Dirichlet boundary condition are proved. The time horizon T is fixed. Making use of the Lions scheme [6], necessary and sufficient conditions of optimality for the Dirichlet problem with the quadratic performance functional with pointwise observation of the state and constrained control are derived.
{"title":"Optimal control of a time delay infinite order parabolic system with pointwise observation of the state","authors":"A. Kowalewski","doi":"10.1109/MMAR.2011.6031322","DOIUrl":"https://doi.org/10.1109/MMAR.2011.6031322","url":null,"abstract":"Various optimization problems for linear parabolic systems with multiple constant time delays are considered. In this paper, we consider an optimal distributed control problem for a linear infinite order parabolic system in which multiple constant time delays appear in the state equation. Sufficient conditions for the existence of a unique solution of the infinite order parabolic time delay equation with the Dirichlet boundary condition are proved. The time horizon T is fixed. Making use of the Lions scheme [6], necessary and sufficient conditions of optimality for the Dirichlet problem with the quadratic performance functional with pointwise observation of the state and constrained control are derived.","PeriodicalId":440376,"journal":{"name":"2011 16th International Conference on Methods & Models in Automation & Robotics","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122956667","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：