Many processes operate only around a limited number of operation points. In order to have adequate control around each operation point, an adaptive controller could be used. Then, if the operation point changes often,...
详细信息
Many processes operate only around a limited number of operation points. In order to have adequate control around each operation point, an adaptive controller could be used. Then, if the operation point changes often, a large number of parameters would have to be adapted over and over again. This prohibits application of conventional adaptive control, which is more suited for processes with slowly changing parameters. Furthermore, continuous adaptation is not always needed or desired. An extension of adaptive control is presented, in which for each operation point the process behaviour can be stored in a memory, retrieved from it and evaluated. These functions are coordinated by a "supervisor". This concept is referred to as supervisory control. It leads to an adaptive control structure which, after a learning phase, quickly adjusts the controller parameters based on retrieval of old information, without the need to fully relcam each time. This approach has been tested on an experimental set-up of a flexible beam, but it is directly applicable to processes in e.g. the (petro)chemical industry as well.
A pulsed priming technique has been applied to series feedback DROs (dielectric resonator oscillators). Using the method, a DRO electrically tuned over 0.4% started in 60 ns without observable starting time jitter, se...
详细信息
A pulsed priming technique has been applied to series feedback DROs (dielectric resonator oscillators). Using the method, a DRO electrically tuned over 0.4% started in 60 ns without observable starting time jitter, settled to with +or-10 ppm in 150 ns, and delivered an output power of 14 dBm with a phase noise level of -92 dBc/Hz. A fixed tuned DRO also showed over 700-ns improvement in starting time when pulse primed with a phase noise level of -95 dBc/Hz. Since resonator Q reduction is not required, the inherent oscillator phase noise is not degraded. The present technique of introducing the priming signal allows for electrical tuning of the oscillator.< >
This survey paper focusses on the main theoretical issues in today’s man-machine systems research and applications. The following problem areas are discussed: (1) modelling human performance and mental workload, with...
详细信息
This survey paper focusses on the main theoretical issues in today’s man-machine systems research and applications. The following problem areas are discussed: (1) modelling human performance and mental workload, with identifying the state of the art as well as major methodological difficulties; (2) task allocation and decision support, with a human-centred perspective on cooperative problem solving, integrated automation, and distributed decision making in teams; (3) man-machine interfaces, with outlining some presentation and dialogue issues; (4) design problems, with stressing the need of early active participation of man-machine-systems specialists and the usefulness of guidelines; and, finally, (5) evaluation and experimental validation, with covering laboratory and field evaluations, with covering laboratory and field evaluations, experimental design and validation, as well as model-driven experimentation. The importance of man-machine-systems contributions to the design of better technical systems and their user acceptability is emphasised.
Many processes operate only around a limited number of operation points. In order to have adequate control around each operation point, an adaptive controller could be used. Then, if the operation point changes often,...
详细信息
Many processes operate only around a limited number of operation points. In order to have adequate control around each operation point, an adaptive controller could be used. Then, if the operation point changes often, a large number of parameters would have to be adapted over and over again. This prohibits application of conventional adaptive control, which is more suited for processes with slowly changing parameters. Furthermore, continuous adaptation is not always needed or desired. An extension of adaptive control is presented, in which for each operation point the process behaviour can be stored in a memory, retrieved from it and evaluated. These functions are coordinated by a “supervisor”. This concept is referred to as supervisory control. It leads to an adaptive control structure which, after a learning phase, quickly adjusts the controller parameters based on retrieval of old information, without the need to fully releam each time. This approach has been tested on an experimental set-up of a flexible beam, but it is directly applicable to processes in e.g. the (petro)chemical industry as well.
Discusses issues relating to the complexity of scientific visualization software system implementation. It is argued that the complexity of current implementations of such systems may limit the utility for users becau...
详细信息
ISBN:
(纸本)9780818628962
Discusses issues relating to the complexity of scientific visualization software system implementation. It is argued that the complexity of current implementations of such systems may limit the utility for users because the interfaces typically require significant knowledge of the data being studied and the applicable visualization algorithms, as well as its infrastructure of graphics, imaging and data handling technology. The issues, unknowns, and possible solutions associated with building effective scientific visualization software are discussed.< >
作者:
Li, K.Tassoudji, M.A.Shin, R.T.Kong, J.A.Department of Electrical Engineering and Computer Science and Research Laboratory of Electronics
Massachusetts Institute of Technology Cambridge Massachusetts 02139 Kevin Li:received his BS and SM degrees (1990) in Electrical Engineering and EE (Electrical Engineer's) degree (1991) from the Massachusetts Institute of Technology
where he is currently pursuing his PhD in Electrical Engineering. Since 1990 he has been the recipient of a United States Air Force Laboratory Graduate Fellowship. His main research interests include electromagnetic scattering radar cross section prediction and numerical techniques. Mr. Li is a member of Tau Beta Pi and a student member of IEEE. M. Ali Tassoudji:received his BS (1987) and MS (1989) degrees in Electrical Engineering from the University of Michigan
Ann Arbor. Since 1989 he has been a research assistant at the Massachusetts Institute of Technology where he is currently working towards his PhD in Electrical Engineering. His research interests include electromagnetic scattering and propagation modeling of microwave circuits and numerical techniques. Mr. Tassoudji is a member of IEEE and Eta Kappa Nu. Robert T. Shin:received his SB (1977)
SM (1980) and PhD (1984) in Electrical Engineering from the Massachusetts Institute of Technology. Since 1984 he has been a member of the Air Defense Techniques Group at MIT Lincoln Laboratory as a Research Staff member and as a Senior Staff member from 1989. His research interests are in the areas of electromagnetic wave scattering and propagation theoretical model development and data interpretation for microwave remote sensing. He is the coauthor ofTheory of Microwave Remote Sensing(Wiley 1985). Dr. Shin is a member of The Electromagnetics Academy IEEE American Geophysical Union Tau Beta Pi Eta Kappa Nu and Commission F of the International Union of Radio Science. Since 1987 he has served on the Editorial Board of theJournal of Electromagnetic Waves and Applications(JEWA). Jin Au Kong:is Professor of Electrical Enginee
The finite difference-time domain (FD-TD) technique is applied to the solution of Maxwell's equations. A computer program, which can be used to simulate and study numerous electromagnetic phenomena, is developed a...
The finite difference-time domain (FD-TD) technique is applied to the solution of Maxwell's equations. A computer program, which can be used to simulate and study numerous electromagnetic phenomena, is developed and implemented on an IBM 386 compatible personal computer. The FD-TD technique is a useful tool for students in electromagnetics. The technique is flexible and can be applied to many basic EM scattering and radiation problems. Because field solutions are found as a function of time, visualization of the propagation of the EM fields is possible. The FD-TD technique is implemented for a two-dimensional rectangular grid in conjunction with a second-order absorbing boundary condition. Both E- and H-field polarizations are analyzed. Finite objects consisting of dielectric, magnetic and conducting materials, and perfectly conducting infinite ground planes are modeled. Plane wave and line current sources are implemented. In addition to the capability of animating the propagation of the EM fields, radiation and scattering patterns can be generated.
The problem of defining an appropriate measure of the degree of nonstationarity for stochastic processes that exhibit cyclostationarity is addressed. After discussing several candidate measures of degree of nonstation...
详细信息
The problem of defining an appropriate measure of the degree of nonstationarity for stochastic processes that exhibit cyclostationarity is addressed. After discussing several candidate measures of degree of nonstationarity, one particularly promising measure is adopted. By decomposing this measure, several component measures are arrived at. Bounds on these measures are derived and their utility in applications involving signal detection and estimation is established. Examples are presented to illustrate the calculation of degrees of nonstationarity for several types of cyclostationary signals.
作者:
P. LöhnbergB. FrankeControl Laboratory
Control Systems and Computer Engineering Group Department of Electrical Engineering University of Twente Enschede The Netherlands
Optimal parameter estimation and experiment design requires a weighting between the cost of parameter errors and that of identification itself. Having identified this need, the general principle to obtain both costs f...
详细信息
Optimal parameter estimation and experiment design requires a weighting between the cost of parameter errors and that of identification itself. Having identified this need, the general principle to obtain both costs for using the model in control design is described. Fast experiment design is achieved by using analytic expressions to calculate the optimal experiment parameters. This is illustrated by a simple example. It is shown that the use of the resulting model criterion in optimal identification is equivalent to cautious stochastic control. A more realistic example is illustrated by simulation.
暂无评论