Search

This paper presents an indoor navigation support system based on the Building Information Models (BIM) paradigm. Although BIM is initially defined for the Architecture, Engineering and Construction/Facility Management (AEC/FM) industry, the authors believe that it can provide added value in this context. To this end, the authors will focus on the Industry Foundation Classes (IFC) standard for the formal representation of BIM. The approach followed in this paper will be based on the ifcOWL ontology, which translates the IFC schemas into Ontology Web Language (OWL). Several modifications of this ontology have been proposed, consisting of the inclusion of new items, SWRL rules and SQWRL searches. This way of expressing the elements of a building can be used to code information that is very useful for navigation, such as the location of elements related to the actions desired by the user. It is important to note that this design is intended to be used as a complement to other well-known tools and techniques for indoor navigation. The proposed modifications have been successfully tested in a variety of simulated and real scenarios. The main limitation of the proposal is the immense amount of information contained in the ifcOWL ontology, which causes difficulties involving its processing and the time necessary to perform operations on it. Those elements that are considered important have been selected, removing those that seem secondary to navigation. This procedure will result in a significant reduction in the storage and semantic processing of the information. Thus, for a system with 1000 individuals (in the ontological sense), the processing time is about 90 s. The authors regard this time as acceptable, since in most cases the tasks involved can be considered part of the system initialization, meaning they will only be executed once at the beginning of the process.

In order to safely navigate populated environments, an autonomous vehicle must be able to detect human shapes using its sensory systems, so that it can properly avoid a collision. In this paper, we introduce a Bayesian approach to the Viola-Jones algorithm, as a method to automatically detect pedestrians in image sequences. We present a probabilistic interpretation of the basic execution of the original tool and develop a technique to produce approximate convolutions of probability matrices with multiple local maxima.

In this paper we consider a recurrent neural network model consisting of two neurons and analyze its stability using the associated characteristic model. In order to analyze the multiple quasi-periodic orbits, the strong resonance of this system, in particular that known as the R2 bifurcation, is also studied. In the case of two neurons, one necessary condition that yields the bifurcation is found. In addition, the direction of the R2 bifurcation is determined by applying normal form theory and the center manifold theorem. The simple conditions for ensuring the existence of multiple quasi-periodic orbits are given. The strong resonance phenomenon is analyzed using numerical simulations and is related with the codimension-two bifurcation of the high-iteration map.

A simple discrete recurrent neural network model is considered. The local stability is analyzed with the associated characteristic model. In order to study the dynamic behavior of the quasi-periodic orbit, it is necessary to determine the Neimark-Sacker bifurcation. In the case of two neurons, one necessary condition that yields the Neimark-Sacker bifurcation is found. In addition to this, the stability and direction of the Neimark-Sacker bifurcation are determined by applying normal form theory and the center manifold theorem. An example is given and a numerical simulation is performed to illustrate the results. The phase-locking phenomena are analyzed for certain experimental results with Arnold Tongues in a particular weight configuration.

In this paper, a study of the odometric system for the autonomous cart Verdino, which is an electric vehicle based on a golf cart, is presented. A mathematical model of the odometric system is derived from cart movement equations, and is used to compute the vehicle position and orientation. The inputs of the system are the odometry encoders, and the model uses the wheels diameter and distance between wheels as parameters. With this model, a least square minimization is made in order to get the nominal best parameters. This model is updated, including a real time wheel diameter measurement improving the accuracy of the results. A neural network model is used in order to learn the odometric model from data. Tests are made using this neural network in several configurations and the results are compared to the mathematical model, showing that the neural network can outperform the first proposed model.

The aim of this paper is to present a knowledge-based approach to supervised classifier design. For this purpose, an expert system has been built following the Common KADS methodology. Classifier design is seen as a general design problem and a modified version of the well-known Propose-Critique-Modify method is proposed as a suitable strategy to solve it. In this context, a number of heuristics are used to shrink the search in the space of possible designs. Although the system is evaluated on different datasets, special emphasis is put on a particular problem: Alzheimer's diagnosis.

A new approach to the classical problem of determining whether or not a set of data has been sampled from a univariate normal distribution is considered. The problem is posed in a pattern recognition framework and the concept of an asymptotically optimal expert is introduced. Some results of Large Deviation theory are used in the selection of experts so that their combined abilities lead to a suitable discrimination procedure. The fact that good asymptotical properties can be extended to finite samples with small sizes is the basis for the rest of the work. The performance of the proposed procedure is also compared to that of some known normality tests. An additional advantage of the procedure is the possibility of computing a reliable measure of support such as the Bayesian posterior probabilities as opposed to the P-values. Furthermore, some extra information can be obtained concerning the type of deviation from normality which is present in the data.

In this study, a procedure planning for a Computer Architecture Course that tries to create a balance between simulation and laboratory procedures is presented. For this planning, we use hardware implementation for parts of this subject that require careful dedication because of its complexity: interconnection networks and parallel processing. On the other hand, we use simulation packages for the teaching of pipelined and vectorial processors. © 2002 Wiley Periodicals, Inc. Comput Appl Eng Educ 10: 11–17, 2002; Published online in Wiley InterScience (www.interscience.wiley.com.); DOI 10.1002/cae.10012

This paper presents a knowledge-based system for aiding in the decision-making process that is carried out in hospital management. There are a number of reasons that have led us to choose a tool such as this one: the amount of information generated in a hospital, its great interrelation and the need of heuristic knowledge for its processing. The KBS has been designed following the KADS methodology. KADS has allowed us to obtain a structured representation of the knowledge, which makes easier both the construction and the debugging of the knowledge base. As a starting point, the decision-making task has been decomposed in four subtasks: monitoring; diagnosis; prediction of the possible solutions for the stated problem; and design of the solution. The prediction task can only be performed through a simulation program where the dynamics of the hospital is modeled. This allows the system to detect the consequences of the application of different possible solutions. The co-operation between simulation and artificial intelligence has proven to be an adequate technique for dealing with the decision-making tasks that are involved with the management of complex organizations.

The objective of our research is to develop computer-based tools to automate the clinical evaluation of the electroencephalogram (EEG) and visual evoked potentials (VEP). This paper describes a set of solutions to support all the aspects regarding the standard procedures of the diagnosis in neurophysiology, including: (1) acquisition and real-time processing and compression of EEG and VEP signals, (2) real-time brain mapping of spectral powers, (3) classifier design, (4) automatic detection of morphologies through supervised neural networks. (5) signal analysis through fuzzy modelling, and (6) a knowledge based approach to classifier design.

The study of a particular complex system by means of computer simulation is described in this paper. Hospitals are chosen as target systems where the proposed methodology is applied.In order to choose the right decisions, hospital managers need all the information about the functioning of the organization. This research project presents a simulation tool that allows virtual societies such as hospitals to be implemented. In this way, the study of emergent behaviours in these systems can be carried out.Petri nets were initially tested as a methodology to model the hospital. However, the complexity of the resulting model led us to follow a process oriented approach.

In this work some optimal control algorithms have been designed for and implemented in a real plant. The plant is a DC motor, controlled in the armature. Both deterministic and stochastic control policies have been developed. The aim of this paper is to show the applicability of optimal control algorithms to improve performances, to look for an empirical proof of the effects that appear in a stochastic policy (such as the caution effect), and to confirm an increment in the performance in the stochastic algorithms as compared with entirely deterministic algorithms. The control schemes applied are Dynamic Programming and Generalized Predictive Control.

Quantitative electroencephalographic (EEG) signal analysis has revealed itself as an important diagnostic tool in the last few years. Through the use of signal processing techniques, new quantitative representations of EEG data are obtained. To automate the diagnosis, a problem of supervised classification must be solved on these. Artificial Neural Networks provide an alternative to more traditional classifier systems for this task. The objective of this paper is to perform a comparison between several classifiers in a particular problem, the brain maturation prediction. The data preprocessing/feature extraction process and the methodology for making the comparison are described. Performance of the methods is evaluated in terms of estimated percentage of correctly classified subjects.

The knowledge acquisition problem is one of the most difficult issues in elaborating a medical expert system. This is more true in the context of automated brain signal diagnosis. This kind of knowledge does not lend itself to be represented in a classical rule-based system and is not easily put in quantitative terms by the specialists. Artificial neural networks (ANNs) provide a useful alternative for capturing this information. In this work, an application of ANNs to brain maturation prediction is presented. The problem is essentially a supervised classification. A case data base consisting of data extracted from electroencephalographic (EEG) signals and diagnoses carried out by an expert neurologist serves to test the ability of several statistical classifiers and several kinds of ANNs in reproducing the expert results. There is also a discussion on how to integrate ANNs in a higher-level knowledge-based system for brain signal interpretation.

The objective of this paper is to study by means of dynamic programming the optimal control of nonlinear continuous systems. We apply to these systems a development of block pulse for the state and a nonuniform discretisation of the state space. As a particular case of a nonlinear system, we analysed a continuous dual control problem, and we carried out an implementation of a stochastic control policy on a real process, a DC motor.

A set of experiments in optimal and predictive control A set of simulation experiments in optimal and predictive control is presented. The paper describes a sequence of eight experiments performed in PASCAL and MATLAB. These experiments have been carried out with the aim of improving the student's understanding of control concepts. A brief discussion of the algorithms and details of each experiment are given.

A simple two-neuron model of a discrete Hopfield neural network is considered. The local stability is analyzed with the associated characteristic model. In order to study the dynamic behavior, the Pitchfork bifurcation is examined. In the case of two neurons, one necessary condition for yielding the Pitchfork bifurcation is found. In addition, the stability and direction of the Pitchfork bifurcation are determined by applying the normal form theory and the center manifold theorem.

This paper presents a set of software tools called that helps in the building knowledge in an autonomous meaningful learning process through an open student model and a student conceptual map explorer. This tool uses adaptive tests based on a Progressive Inquiry (PI) model and has been used for teaching Computer Architecture (in particular the domain of Computer memory hierarchy) although the proposed system is supposed to be valid for any knowledge domain. The mentioned field has been chosen due to its limited complexity but it is sufficiently general to be learnt in several disciplines with different levels of detail.

The goal of this work is the development of an intelligent tool that would help teachers and students in order to improve the learning process, making use of a variation of the concept map paradigm. The authors propose a novelty in their use for determining the actual student status, that is, what a student currently knows, which concepts a student needs to revise, etc.

When computing the trajectory of an autonomous vehicle, inevitable collision states must be avoided at all costs, so no harm comes to the device or pedestrians around it. In dynamic environments, considering collisions as binary events is risky and inefficient, as the future position of moving obstacles is unknown. We introduce a time-dependent probabilistic collision state checker system, which traces a safe route with a minimum collision probability for a robot. We apply a sequential Bayesian model to calculate approximate predictions of the movement patterns of the obstacles, and define a time-dependent variation of the Dijkstra algorithm to compute statistically safe trajectories through a crowded area. We prove the efficiency of our methods through experimentation, using a self-guided robotic device.

In this paper we present two methods based on neural networks (NN) for resolution of nonlinear systems optimal control with arbitrary performance index. We have used the minimum time index as an example. Both methods solve the optimal problem for a region of the state space by means of a multistage optimization through a NN chain. Each NN has a fully connected feedforward multilayer structure and the training algorithm for the NN chain is the backpropagation. The chain structure is different for each method, as well as the discretization procedure: classical and block pulse function. >

SICOLE is a research effort to develop a software environment to help tutors of dyslexic children with the diagnostic and treatment tasks. This paper describes the architecture of the package already implemented where three important elements interact: a multimedia interface, an inference module and a database. This architecture provides the system with the flexibility to support a large variety of tasks, dynamic presentations, and complex teaching strategies. The application of speech recognition technology is also researched as an important part of the evaluation of the dyslexic children improvements. This package is being used at the present moment in several Spanish Schools as part of its validation process.

The objective of our research is to develop computer-based tools to automate the clinical evaluation of the electroencephalogram (EEG) and visual evoked potentials (VEP). This paper describes a set of solutions to support all the aspects regarding the standard procedures of the diagnosis in neurophysiology, including: (1) acquisition and real-time processing and compression of EEG and VEP signals, (2) real-time brain mapping of spectral powers, (3) classifier design, (4) automatic detection of morphologies through supervised neural networks. (5) signal analysis through fuzzy modelling, and (6) a knowledge based approach to classifier design.

A simple class of discrete recurrent neural network is analyzed to establish its possible chaotic dynamics. A two-neuron network is selected for simplicity and two cases of chaotic dynamics are distinguished, one degenerate (one-dimensional) and a more general situation with a two-dimensional attractor. The robustness of the found configurations is assessed through evaluation of Lyapunov exponents in a range of parameter values. In every situation there is a lack of robustness, as suggested in a conjecture of Barreto et al.

This paper presents a set of methods for helping in the analysis of signals with particular features that admit a symbolic description. The methodology is based on a general discrete model for a symbolic processing subsystem, which is fuzzyfied by means of a fuzzy inference system. In this framework a number of design problems have been approached. The curse of dimensionality problem and the specification of adequate membership functions are the main ones. In addition, other strategies, which make the design process simpler and more robust, are introduced. Their goals are automating the production of the rule base of the fuzzy system and composing complex systems from simpler subsystems under symbolic constrains. These techniques are applied to the analysis of wakefulness episodes in the sleep EEG. In order to solve the practical difficulty of finding remarkable situations from the outputs of the symbolic subsystems an unsupervised adaptive learning technique (FART network) has been applied.

The objective of this study is to evaluate cerebral maturity by means of quantitative analysis techniques applied to the electroencephalogram (EEG).A transversal study of cerebral maturity was carried out in 403 persons who had undergone an EEG. A previous pilot study had been carried out of 103 persons. A series of spectral parameters of the EEG were selected so that all those studied were in the most similar conditions possible. Different frequency bands were analyzed choosing the one with best discrimination of the maturity aspect. Classification of the different levels of cerebral maturity was done with the help of multivariant analysis. The value of the median of the frequencies and the spectrum of relative potencies at the moment when a frequency band is at a maximum are the parameters which evolve best with age and best discriminate maturity Spectral analysis allows selection of the frequency bands most suitable to the problem. Working with two frequency bands is sufficient to evaluate cerebral maturity.The variables obtained in the occipital channels were sufficient for evaluation of cerebral maturity. Those of the right hemisphere were more significant for diagnosis. The occipital channels are the most relevant in the study of cerebral maturity.The neuronal network is the most efficient classifier for classification of different groups of maturity The next most efficient method is by quadratic discriminant analysis. Consideration of the variables, taking into account the factors of stability and regularity of the EEG signals improves discrimination with respect to the average of those recorded during the entire procedure.

The aim of this job is to evaluate brain maturation by means of Electroencephalogram (EEG) and Visual Evoked Potentials stimulated with flash (VEP-flash) quantitative analysis techniques.The transversal study is made on a sample of 96 subjects in which EEG and VEP-flash, first isolated and then joining both, are analyzed. The selection of spectral parameters was done taking care of all the subjects were selected in the sense of maximizing brain maturation discrimination. Multivariate analysis techniques for classifying subjects were used. EEG and VEP-flash variables were selected with the linear discriminant analysis. In the EEG case the variables take into account, as a reference, either the median of the power spectrum or either the time instant in which the spectral power in every band reaches its maximum value. In the joined EEG-VEP-flash the VEP variables which give more information were related with the slopes and distances between the basic peaks of the evoked response (N1, P1 and N2) and age. For brain maturation evaluation the variables in the occipital channels are sufficient, being those of the right hemisphere the most diagnostic significative ones.The joined use of EEG and VEP-flash means an improvement in the maturative level discrimination regarding to the isolated consideration of any of them. Variables obtained from the EEG-VEP-flash are enough for brain maturation evaluation.

An application of Neural Networks (NN) to brain maturation prediction is presented. The problem consists of, given a pattern extracted from electroencephalographic (EEG) signals, state the degree of brain development (low or normal/high). To that end, a population of subjects with their EEG assessed by a neurologist is available. A Backpropagation (BP) neural network is used for this supervised classification task, and a comparison with standard statistical classifiers is made. The effect on performance of several preprocessing techniques such as Principal Components Analysis (PCA), normalization and scaling is investigated. It is found better performance in the NN approach, both in terms of efficiency and consistency.

A collection of practical experiments in the area of multirate digital signal processing and filter banks is described. The systems are implemented using a block oriented simulation tool together with specific software. The attention is focused on the benefits that simulation gives in understanding the main topics related with the multirate designs.

Computer simulation has eased the study of complex systems. A hospital is a complex sys tem that is formed by a large number of units with strong interrelationships. Even though resources are limited, patients must be effi ciently treated. This paper presents simulation as a tool to aid hospital management. In the first phase we present an introduction to the problem and its motivation. The next step is the description of how the system functions. The choice of the simulation model and the approach in dealing with it are described. Finally, the implementation of the simulation tool is pre sented. This tool is used for supporting the deci sion-making processes in hospital management.

The fundamental objective is the development of an automated system for aiding in the diagnosis of brain signals. The authors' aim is to isolate the end user from the multidisciplinary knowledge integrated in the system. To this end, it has been designed in several layers or levels of abstraction. The bottom layers mainly comprise procedural knowledge and in the upper layers one finds symbolic knowledge, closer to the expert neurophysiologist terms. Two examples are introduced, showing the different nature of the several kinds of problems in this field, in addition to revealing the methodologies employed in their resolution. Finally, the authors describe the general lines of the design and implementation of the proposed system.

The objective of our research is to develop computer based tools to automate the clinical evaluation of the electroencephalogram (EEG) and visual evoked potentials (VEP). The paper describes a set of solutions to support all the aspects regarding the standard procedures of the diagnosis in neurophysiology. More specifically, the paper focuses on the description of an approach to automate the process carried out by the physician when an EEG register has to be used to obtain a diagnostic of a patient. Two main lines have been established. The first one deals with the development of support tools to aid in the technical aspects of the process: data acquisition, data compression and brain map representation. The second line is focused on the classification and learning of neurophysiological signals. A classical approach has been successfully applied to the problem of brain maturation classification.

Experiments on a d.c. motor-based system for a digital control course A set of real-time experiments is presented. These experiments are implemented on a position-velocity system and cover a wide range of control techniques. We start with basic control actions and continue with more complex strategies like optimal control. We emphasize the study of all practical aspects of the experiences: noise, delays, etc.

The study of a particular complex system by means of computer simulation is described in this paper. Hospitals are chosen as target systems where the proposed methodology is applied. In order to choose the right decisions, hospital managers need all the information about the functioning of the organization. This research project presents a simulation tool that allows virtual societies such as hospitals to be implemented. In this way, the study of emergent behaviors in these systems can be carried out. The methodology used to model the hospital is process oriented. This approach allows us to implement a patient-centered simulation tool.

The aim of this work is to provide a new approach to the classical problem of determining whether or not a set of data has been sampled from a univariate normal distribution. A simple neural network architecture is proposed as an efficient way of combining and reinforcing the discriminatory capabilities of different popular statistics commonly used in conventional hypothesis testing procedures. Special emphasis is placed on the fact that these procedures lack a reliable measure of the degree to which the observed data supports the normality assumption. Several authors have shown that the so-called P-values are inefficient and ambiguous when dealing with this matter, so the Bayesian posterior probabilities are suggested as the best candidates to play this role. For this reason, a significant part of our work has focused on training the neural networks so that their outputs accurately approximate these probabilities.