Current computer aided control system design environments seldom support optimization methods for controller design in a truly interactive manner. A prototype tool, called PARSIM, supporting parallel processing, optim...
详细信息
Current computer aided control system design environments seldom support optimization methods for controller design in a truly interactive manner. A prototype tool, called PARSIM, supporting parallel processing, optimization and a graphical user interface is presented addressing many of the problems inherent in current approaches to multiobjective optimization based design methods. An XWindows interface is used to simplify problem formulation and control the optimization processes. Using a previously developed interface, the computational burden may be alleviated by parallel processing.< >
Most speech recognition systems are unable to cope with data from high-resolution pre-processors (such as auditory models and high-resolution spectral estimates) for two reasons. One is due to the inappropriateness of...
详细信息
Most speech recognition systems are unable to cope with data from high-resolution pre-processors (such as auditory models and high-resolution spectral estimates) for two reasons. One is due to the inappropriateness of measures related to the Euclidean distance. The other is somewhat less obvious, but is due to the non-ergodic nature of short-term parameterisations of speech sounds. This aspect of speech variability is addressed. The authors show how a linear, but nonstationary, vector predictor, based on the concept of 'acoustic flow', can be used to estimate the redundancy in speech data, paving the way for an improvement in recognition performance.< >
The structure of multivariable systems is examined and the derivation of multivariable models using experimental test procedures is explored. An example of a laboratory scale multivariable plant, a motor/alternator ri...
详细信息
The structure of multivariable systems is examined and the derivation of multivariable models using experimental test procedures is explored. An example of a laboratory scale multivariable plant, a motor/alternator rig, is described and results are presented to illustrate how simple transfer function models can be derived from step response tests. The design of decoupling precompensators, are shown to be useful in removing interactions. The basic features of a second multivariable laboratory-scale process with different dynamic characteristics are also included.
The general trend in muscle relaxant anaesthesia is for shorter acting agents which allow the patient to recover from paralysis quickly with minimal use of reversal agent. This necessitates continuous infusion schemes...
详细信息
The general trend in muscle relaxant anaesthesia is for shorter acting agents which allow the patient to recover from paralysis quickly with minimal use of reversal agent. This necessitates continuous infusion schemes which must accommodate a wide range of patient characteristics. This paper investigates the application of self-organizing fuzzy logic control (SOFLC) to automated infusion of atracurium. This scheme offers a simple means to develop a nonlinear controller which quickly establishes the level of muscle relaxant required by each individual patient in a self-learning fashion. Preliminary results show this approach is suitable for clinical application.
A direct inverse neural network-based control scheme is presented and its performance tested by computer simulations using a nonlinear system. The adaptation of the neural network weights involved in the control schem...
详细信息
A direct inverse neural network-based control scheme is presented and its performance tested by computer simulations using a nonlinear system. The adaptation of the neural network weights involved in the control scheme is carried out by means of a continuous-time variable structure control-based learning algorithm that enables an online approximation of the inverse transfer operator of the unknown system to be obtained. Further research is underway to prove conditions for assuring the stability of the proposed control scheme.< >
Al planning systems must be able to update their internal model of the domain in order to determine the effects of a possible action. The most widely used update method is the STRIPS operator. STRIPS operators are ext...
详细信息
Complex and computationally demanding algorithms are often employed to satisfy the performance requirements of modern controlsystems. Conventional uni-processor and homogeneous parallel architectures can not meet the...
详细信息
Complex and computationally demanding algorithms are often employed to satisfy the performance requirements of modern controlsystems. Conventional uni-processor and homogeneous parallel architectures can not meet the performance requirements of these algorithms. A suitable solution maybe obtained with a computationally specialised heterogeneous architecture. This paper introduces the key issues associated with the development of this class of systems. followed by a description of a set of design tools used to automate the development. One of the most challenging aspects of the system development is the scheduling of tasks to processors. This paper will focus closely on this issue by discussing the results of a variety of scheduling heuristics used within the design tools.
The performance of protocols currently used for controlling the infusion of propofol intravenously and which assume a model based on an average pharmacokinetics population to predict the plasma concentration are shown...
详细信息
The performance of protocols currently used for controlling the infusion of propofol intravenously and which assume a model based on an average pharmacokinetics population to predict the plasma concentration are shown to degrade when mismatch conditions between the patient and the model are considered. The pharmacokinetic model hence considered was extended to include pharmacodynamics by considering Mean Arterial Pressure measurements (thought to give a good indication of the anaesthetic state). Using this new extended model, the performances of two control strategies, modified Alvis algorithm and fixed Generalised Predictive control (GPC) algorithm were assessed. The study reveals that under mismatch conditions the fixed GPC controller displays better properties than the modified Alvis algorithm making it a more attractive candidate for future clinical trials in theatre.
This note is concerned with structure selection of nonlinear models. By structure selection it is meant the choice of the model basis prior to parameter estimation. It is argued that the effect of the sampling period ...
This note is concerned with structure selection of nonlinear models. By structure selection it is meant the choice of the model basis prior to parameter estimation. It is argued that the effect of the sampling period and the noise on the data may in some cases preclude adequate structure selection. Other issues which are briefly discussed include the selection of the sampling period and the effectiveness of information criteria in indicating the best size of a nonlinear model.
暂无评论