The present paper proposes a novel technique for improving the performance of a prototype chaotic communication system by increasing the efficiency of the chaos-control algorithm. With the availability of such chaos-c...
详细信息
The present paper proposes a novel technique for improving the performance of a prototype chaotic communication system by increasing the efficiency of the chaos-control algorithm. With the availability of such chaos-controlalgorithms it is appropriate to consider practical devices in which chaos can be easily generated and hence utilised in communications systems and in particular optical communications systems.< >
The paper presents an investigation into the utilisation of parallel processing techniques in the real-time simulation of a flexible manipulator system. The performance demands of modern controlsystems require the em...
详细信息
The paper presents an investigation into the utilisation of parallel processing techniques in the real-time simulation of a flexible manipulator system. The performance demands of modern controlsystems require the employment of complex algorithms with demanding operations. This, in turn, leads to shorter sampling times. Therefore, real-time performances with complex algorithms become difficult to achieve. Many demanding complex processes cannot be satisfactorily realised with conventional uni-processor and multi-processor systems. Thus, alternative strategies where multi-processor based systems are employed, making effective use of parallel processing techniques, could provide suitable solutions. A finite dimensional simulation of a single-link flexible manipulator system is developed through discretisation of the governing dynamic equation of the system both in time and space co-ordinates using finite difference approximation methods. The proposed algorithm allows for dynamic modification of the boundary conditions and inclusion of distributed actuator and sensor terms in the system dynamic equation.< >
The authors present a practical solution to the problem of real-time robot control including the nonlinear dynamic model of the manipulator by employing a parallel processing approach. The parallelism inherent in the ...
详细信息
The authors present a practical solution to the problem of real-time robot control including the nonlinear dynamic model of the manipulator by employing a parallel processing approach. The parallelism inherent in the adaptive controllers is exploited to obtain an efficient implementation that reduces the overall computation time to within the limit acceptable for real-time control. The distributed algorithm is implemented on a network of transputers for the six-joint PUMA 560 arm.< >
Neural and fuzzy algorithms have been introduced as basic adaptive nonlinear systems, therefore it is only natural that control engineers should investigate these models in order to determine whether or not they form ...
详细信息
Neural and fuzzy algorithms have been introduced as basic adaptive nonlinear systems, therefore it is only natural that control engineers should investigate these models in order to determine whether or not they form a useful paradigm which can be applied in certain cases. They are not a universal panacea for all modelling and control problems and just because they learn to form nonlinear mappings does not necessarily mean that they will outperform more conventional nonlinear modelling techniques. From the engineering community, research should be aimed at establishing the modelling and learning abilities of these adaptive neural and fuzzy systems in order to determine whether they will form another methodology which can be applied to real-world problems; irrespective of the fact that these algorithm may have some vague biological relevance. In this paper the authors investigate the modelling and generalisation abilities of a certain class of neural algorithms called neurofuzzy networks.< >
The learning algorithms and structures which have become popular in the neural network community over the last few years have been successfully applied to various challenging modelling problems. In contrast to the lin...
详细信息
The learning algorithms and structures which have become popular in the neural network community over the last few years have been successfully applied to various challenging modelling problems. In contrast to the linear modelling methods, there is still no clearly defined engineering process from which a model can reliably be created from measurements taken from a physical system. Problems for the practical application of neural networks involve the interpretation of the trained models, and the explicit introduction of a priori models into the learning system, as well as the use, in many cases, only basic ad hoc validation and experiment design techniques. This talk will discuss several methods which can improve the engineering aspects of model identification with neural nets.< >
The authors present an investigation into the utilisation of parallel computing techinques for real-time simulation and control of a flexible beam structure in transverse vibration. The performance demands of modern c...
详细信息
The authors present an investigation into the utilisation of parallel computing techinques for real-time simulation and control of a flexible beam structure in transverse vibration. The performance demands of modern controlsystems require the employment of complex algorithms with demanding operations which, in turn, leads to shorter sampling times. Therefore, real-time performance in control applications where the use of advanced control methods is warranted becomes difficult to accomplish. Many demanding complex control processes cannot be satisfactorily realised with conventional uni-processor and multi-processor systems. Previous investigations have demonstrated the limitations of employing only transputers for real-time implementations in control applications. Alternative strategies where multi-processor based systems are employed, utilising digital signal processing (DSP) and parallel processing techniques, could provide suitable methodologies.< >
The study presented in this paper is centred around a VME based multiprocessor, multi-axis control system, which has wide application in high precision production machines. A mechatronics approach adopted throughout t...
详细信息
The study presented in this paper is centred around a VME based multiprocessor, multi-axis control system, which has wide application in high precision production machines. A mechatronics approach adopted throughout the project, which brought together many technologies, has resulted in a very dramatic improvement in system performance. The key to this improvement was finding a control strategy to reshape the overall system dynamics resulting in a much higher bandwidth and improved dynamic response. This programme of research has demonstrated that state control techniques offer a viable approach to designing distributed intelligent controls for fast acting precision servo systems. Two self commissioning strategies were implemented, the results are discussed and compared to the 'tuning parameters' set by an experienced commissioning engineer.< >
One of the important problems to be solved for neural network applications is to find a suitable network structure solving the given task. To reduce the engineering efforts for the architecture design a data driven al...
详细信息
One of the important problems to be solved for neural network applications is to find a suitable network structure solving the given task. To reduce the engineering efforts for the architecture design a data driven algorithm is desirable which constructs a network structure during the learning process. There are different approaches for structure adaptation with evolutionary algorithms, growth algorithms and others. To solve large problems successfully it is necessary to divide the problem into subproblems and to solve them separately by experts. This is a fundamental principle of nature. To implement this principle in artificial neural networks there are different approaches, but these algorithms yield fixed network structures. The authors propose a learning architecture for growing complex artificial neural networks which tries to include both sides of the coin, structure adaptation and task decomposition. The growing process is controlled by self-observation or reflexion. The algorithm generates a feedforward network bottom up by cyclically inserting cascaded hidden layers. Inputs of a hidden layer unit are locally restricted with respect to the input space by using a new kind of activation function, combining the local characteristics of radial basis function units with sigmoid units. Contrary to the cascade-correlation learning architecture the authors introduce different correlation measures to train the network units featuring different goals. The task decomposition between subnetworks is done by maximizing the anticorrelation between the hidden layer units output and a connection routing algorithm which only connects cooperative units of different layers. These features resemble the TACOMA (TAsk decomposition, COrrelation Measures and local Attention neurons) learning architecture. Self-observation is done by transforming the errors and the network structure to the input space. So it is possible to infer from errors to structure and reverse.< >
geneticalgorithms (GA) are adaptive search techniques, based on the principles of natural genetics and natural selection, which, in controlsystemsengineering, can be used as an optimization tool or as the basis of ...
详细信息
geneticalgorithms (GA) are adaptive search techniques, based on the principles of natural genetics and natural selection, which, in controlsystemsengineering, can be used as an optimization tool or as the basis of more general adaptive systems. Following an introduction to the simple GA, important characteristics of GA are identified and control applications are described.< >
Multiobjective geneticalgorithms (MOGAs) are introduced as a modification of the standard genetic algorithm at the selection level. Rank-based fitness assignment and the implementation of sharing in the objective val...
详细信息
Multiobjective geneticalgorithms (MOGAs) are introduced as a modification of the standard genetic algorithm at the selection level. Rank-based fitness assignment and the implementation of sharing in the objective value domain are two of the important aspects of this class of algorithms. The ability of the decision maker (DM) to progressively articulate its preferences while learning about the problem under consideration is one of their most attractive features. Illustrative results of how the DM can interact with the genetic algorithm are presented. They also show the ability of the MOGA to uniformly sample regions of the trade-off surface.< >
暂无评论