Abstract: In order to control systems that contain nonlinearities or uncertainties, control strategies must deal with the effects of these Since most control methods based on mathematical models have been mainly focused on stability robustness against nonlinearities or uncertainties, they are limited in their ability to improve the transient responses In this paper, a nonlinear fuzzy PID control method is suggested, which can stably improve the transient responses of systems disturbed by nonlinearities or unknown mathematical characteristics Although the derivation of the control law is based on the design procedure for general fuzzy logic controllers, the resultant control algorithm has analytical form with time varying PID gains rather than linguistic form This means the implementation of the proposed method can be easily and effectively applied to real-time control situations Control simulations are carried out to evaluate the transient performance of the suggested method through example systems, by comparing its responses with those of the nonlinear fuzzy PI control method developed in [9]

Abstract: In this chapter we discuss our research on multimodal interaction in a virtual environment. The environment we have developed can be considered as a ‘laboratory’ for research on multimodal interactions and multimedia presentation, where we have multiple users and various agents that help the users to obtain and communicate information. The environment represents a theatre. The theatre has been built using VRML (Virtual Reality Modeling Language) and it can be accessed through World Wide Web (WWW). This virtual theatre allows navigation input through keyboard function keys and mouse, but there is also a navigation agent which tries to understand keyboard natural language input and spoken commands. Feedback of the system is given using speech synthesis. We also have Karen, an information agent which allows a natural language dialogue with the user. In development are several talking faces for the different agents in the virtual world. We investigate how we can increase the user’s commitment to the environment and its agents by providing context and increasing the user’s feeling of ‘presence’ in the environment.

Abstract: Notions of similarity and neighborhood play an important role in informatics. Different disciplines have developed their own treatment of related measures. We consider this problem under the viewpoint of case based reasoning and fuzzy theory. While distance and similarity can be considered to be formally equivalent, there exist some differences concerning their intuitive use which have impact on the composition of global measures from local ones.

Abstract: Interest in issues relating to consciousness has grown markedly during the last several years. And yet, nobody can claim that consciousness is a well-understood concept that lends itself to precise analysis. It may be argued that, as a concept, consciousness is much too complex to fit into the conceptual structure of existing theories based on Aristotelian logic and probability theory. An approach suggested in this paper links consciousness to perceptions and perceptions to their descriptors in a natural language. In this way, those aspects of consciousness which relate to reasoning and concept formation are linked to what is referred to as the methodology of computing with words (CW). Computing, in its usual sense, is centered on manipulation of numbers and symbols. In contrast, computing with words, or CW for short, is a methodology in which the objects of computation are words and propositions drawn from a natural language (e.g., small, large, far, heavy, not very likely, the price of gas is low and declining, Berkeley is near San Francisco, it is very unlikely that there will be a significant increase in the price of oil in the near future, etc.). Computing with words is inspired by the remarkable human capability to perform a wide variety of physical and mental tasks without any measurements and any computations. Familiar examples of such tasks are parking a car, driving in heavy traffic, playing golf, riding a bicycle, understanding speech, and summarizing a story. Underlying this remarkable capability is the brain's crucial ability to manipulate perceptions--perceptions of distance, size, weight, color, speed, time, direction, force, number, truth, likelihood, and other characteristics of physical and mental objects. Manipulation of perceptions plays a key role in human recognition, decision and execution processes. As a methodology, computing with words provides a foundation for a computational theory of perceptions: a theory which may have an important bearing on how humans make--and machines might make--perception-based rational decisions in an environment of imprecision, uncertainty, and partial truth. A basic difference between perceptions and measurements is that, in general, measurements are crisp, whereas perceptions are fuzzy. One of the fundamental aims of science has been and continues to be that of progressing from perceptions to measurements. Pursuit of this aim has led to brilliant successes. We have sent men to the moon; we can build computers that are capable of performing billions of computations per second; we have constructed telescopes that can explore the far reaches of the universe; and we can date the age of rocks that are millions of years old. But alongside the brilliant successes stand conspicuous underachievements and outright failures. We cannot build robots that can move with the agility of animals or humans; we cannot automate driving in heavy traffic; we cannot translate from one language to another at the level of a human interpreter; we cannot create programs that can summarize non-trivial stories; our ability to model the behavior of economic systems leaves much to be desired; and we cannot build machines that can compete with children in the performance of a wide variety of physical and cognitive tasks. It may be argued that underlying the underachievements and failures is the unavailability of a methodology for reasoning and computing with perceptions rather than measurements. An outline of such a methodology--referred to as a computational theory of perceptions--is presented in this paper. The computational theory of perceptions (CTP) is based on the methodology of CW. In CTP, words play the role of labels of perceptions, and, more generally, perceptions are expressed as propositions in a natural language. CW-based techniques are employed to translate propositions expressed in a natural language into what is called the Generalized Constraint Language (GCL). In this language, the meaning of a proposition is expressed as a generalized constraint, X isr R, where X is the constrained variable, R is the constraining relation, and isr is a variable copula in which r is an indexing variable whose value defines the way in which R constrains X. Among the basic types of constraints are possibilistic, veristic, probabilistic, random set, Pawlak set, fuzzy graph, and usuality. The wide variety of constraints in GCL makes GCL a much more expressive language than the language of predicate logic. In CW, the initial and terminal data sets, IDS and TDS, are assumed to consist of propositions expressed in a natural language. These propositions are translated, respectively, into antecedent and consequent constraints. Consequent constraints are derived from antecedent constraints through the use of rules of constraint propagation. The principal constraint propagation rule is the generalized extension principle. (ABSTRACT TRUNCATED)

Abstract: A new EMG pattern classification method based on soft computing techniques is proposed to help the disabled and the elderly handle rehabilitation robotic arm systems. First, it is shown that EMG is more useful than existing input devices, such as voice, a laser pointer, and a keypad in view of naturality, extensibility, and applicability. Next, through soft computing techniques, such as the fuzzy logic and rough set theory, a new procedure is proposed to select an essential feature set of EMG signals that is independent of users. In order to classify pre-defined motions, a fuzzy pattern classification and fuzzy min-max neural networks (FMMNN) are adopted to handle the selected minimal feature set in systematical ways. As results, motions are recognized with success rates of 83 percent and 90 percent for fuzzy pattern classification and FMMNN, respectively.

Abstract: The article presents an attempt to apply an artificial neural network in anti-collision systems at sea. Such systems determine safe trajectory for a ship by, most frequently, processing the information sourced from an automatic radar plotting aids (ARPA) type arrangement. In this research work the use of neural classifier has been proposed as an element supporting a navigator in the process of determining the ship’s domain, i.e., the area around a ship which, for safety reasons, should remain free from navigational obstacles. Neural classifier has a form of a multilayer feed-forward network of a perceptronic nature. It has been assigned a task to represent the evaluation of a navigational situation provided by an experienced navigator functioning as a teacher in the process of the network learning phase. Programs designed in the MATLAB language have been applied for the simulation of the network and also for the illustration of a navigational situation. Testing of the correctness of classification of the collision situations by the network have been also conducted. In the final part of the article conclusions have been formulated with regard to the application of the neural classifier in the process of determining a safe trajectory of a ship.

Abstract: In this paper, we propose a robust evolutionary algorithm, called adaptive mutations genetic algorithm, for function optimization problems. Our main contribution is robustly optimizing problems whose number of variables from 2 to 200. In order to have a fair comparison, we propose the criteria for constructing a testing bed and for classifying these problems into different complexity degrees. The proposed approach, based on the family competition and multiple adaptive rules, successfully integrates the decreasing-based Gaussian mutation and self-adaptive Cauchy mutation to balance the exploitation and exploration. It is implemented and applied to widely used test functions and several nonseparable multimodal functions. Experimental results indicate that our approach is more robust than ten evolutionary algorithms.

Abstract: This paper illustrates how internal model control of nonlinear processes can be achieved by recurrent neural networks, e.g. fully connected Hopfield networks. It is shown that using results developed by Kambhampati et al. (1995), that once a recurrent network model of a nonlinear system has been produced, a controller can be produced which consists of the network comprising the inverse of the model and a filter. Thus, the network providing control for the nonlinear system does not require any training after it has been trained to model the nonlinear system. Stability and other issues of importance for nonlinear control systems are also discussed.

Abstract: The concept of an image of a fuzzy set under a fuzzy relation has proved to be a very powerful tool in fuzzy set theoretical applications. In this paper, we explain how it can be used to model linguistic expressions. For the representation of expressions, such as “at least middle-aged”, “brighter than average”, we will use fuzzy ordering relations, while resemblance relations will be suitable to model linguistic terms, such as “more or less expensive” and “very tall.” We will show how these representations can be smoothly integrated in approximate reasoning schemes using the compositional rule of inference.

Abstract: The ORBEX coprocessor has been designed to execute the typical fuzzy operations of a system based on fuzzy rules. The first real application has been fuzzy controllers for electric cars. The values of the input variables, the position and the orientation of the car with respect the desired trajectory of reference, are obtained from the data provided by a DGPS boarded in the vehicle. The values of the output variables provided by the controller are the angle that the steering wheel has to be turned and the increment of the velocity. Due to the power of the ORBEX coprocessor the controllers can be written with few though meaningful rules.

Abstract: The Self-Organizing Map (SOM) is one of the most popular neural network methods. It is a powerful tool in visualization and analysis of high-dimensional data in various engineering applications. The SOM maps the data on a two-dimensional grid which may be used as a base for various kinds of visual approaches for clustering, correlation and novelty detection. In this chapter, we present novel methods that enhance the SOM based visualization in correlation hunting and novelty detection. These methods are applied to two industrial case studies: analysis of hot rolling of steel and continuous pulp process. A research software for fast development of SOM based tools is brieey described.

Abstract: This paper introduces a novel distributed fault recovery technique that could be useful in the design of re-configurable holonic manufacturing systems. The method minimizes the fuzzy entropy in the vague information spread across the multi-agent system (MAS). Minimal entropy induces optimal grouping of the agents in holonic clusters. After a brief introduction to the multi-agent modeling of holonic manufacturing systems, we present a framework for distributed control of holonic re-configurable production systems. Then our approach to fuzzy modeling of MASs is exposed as a tool enabling emergent holonic clustering of the agents. Implemented at the logical level of resource grouping in our distributed control mechanism, the proposed approach enables system re-configuration. The emerging new structures can accomplish fault-recovery by re-distribution of tasks in case a resource fails and as well can re-configure production by dynamic re-allocation of resources for newly received high-priority tasks.

Abstract: We present a cooperative, multistep negotiation mechanism for multiagent systems. This mechanism uses marginal utility gain and marginal utility cost to structure the negotiation process. This enables an agent to understand another agent’s situation in order to find a solution that increases their combined utility. These two values summarize the agent’s local information and reduce the communication load. We also introduce a multiple attribute utility theory into negotiations. This allows agents to negotiate over multiple attributes of the commitment which makes it more likely for agents to find a solution that increases the global utility by producing more options.

Abstract: The Intelligent Assistant (IA) is an integrated system of intelligent software agents that helps the user with communication, information and time management. The IA includes specialist assistants for e-mail prioritization and telephone call filtering (communication management), Web search and Yellow Pages® lookup (information management), and calendar scheduling (time management). Each such assistant is designed to have a model of the user and a learning module for acquiring user preferences. In addition, the IA includes a toolbar providing a graphical interface to the system, a multimodal interface for accepting spoken commands and tracking the user’s activity, and a co-ordinator responsible for managing communication from the system to the user and for initiating system activities on the user’s behalf. A primary design objective of the IA is that its operation is as transparent as possible, to enable the user to control the system as far as is practicable without incurring a heavy overhead when creating and modifying the system’s behaviour. Hence each specialist assistant is designed to represent its user model in a way that is intuitively understandable to non-technical users, and is configured to adaptively modify its user model through time to accommodate the user’s changing preferences. However, in contrast to adaptive interface agents built under the behaviour-based paradigm, the assistants in the IA embrace complex AI representations and machine learning techniques to accomplish more sophisticated behaviour.

Abstract: The paper contains a review of some results concerning probability theory on MV algebras (laws of large numbers, central limit theorem, martingale convergence theorem). Also some algebraic and methodical aspects are discussed.

Abstract: The present work deals with the development of simple finite element (FE) models of a turbocharger (rotor, foundation and hydrodynamic bearings) combined with neural networks and identification methods and vibration data obtained from real machines towards the automatic fault diagnosis. The development of this system is based on four sequential steps: the first is the development of simple but realistic FE models based on dynamic simulations of the complete system. The second step is the monitoring of the real turbocharger. The third step is the accurate modelling of the foundations and the excitation from the main engine, which will be done using a robust optimisation method. In the fourth step all the possible faults of the machine are identified using the artificial neural networks (ANN). In this way we can take advantage of the ANN learning capability for the real time diagnosis of potential faults. The application of the proposed system to a real naval turbocharger with vibration data obtained on working conditions show some promising results.

Abstract: Although the fuzzy retrieval model constitutes a powerful extension of the Boolean one, being able to deal with the imprecision and subjectivity existing in the Information Retrieval process, users are not usually able to express their query requirements in the form of an extended Boolean query including weights. To solve this problem, different tools to assist the user in the query formulation have been proposed. In this paper, the genetic algorithm-programming technique is considered to build an algorithm of this kind that will be able to automatically learn weighted queries -modeling the user's needs- for a fuzzy information retrieval system by applying an off-line adaptive process starting from a set of relevant documents.

Abstract: Computer vision inspection systems are widely used for on-line inspection and quality control to improve the finished product quality and lower the costs. In this study, a novel system for wineglass defect inspection has been presented. A carrier grating is used to carry spatial information about inspected glasses. To demodulate the regular/irregular information, an analysis approach for extracting texture features based on Gabor filter has been proposed. From global response surfaces or feature images, objects of interest are segmented. Next local response surfaces or feature images over the isolated regions are calculated. Lastly texture features describing their smoothness and uniformity are computed based on a gray-level co-occurrence matrix and extracted from the local feature images. To make an acceptance/rejection decision, these features are fed to a conventional back propagation classifier. The proposed inspection system has been implemented and tested with some wineglass samples.

Abstract: This chapter presents a motion detector for interlaced video based on the principles of fuzzy logic control. Interlaced video exhibits several possible artifacts such as line flicker and line crawling. De-interlacing algorithms convert interlaced to progressive scan video formats by interpolating the missing lines. These algorithms can be used to improve picture quality but are also necessary to display interlaced video on progressive output devices. A good de-interlacing technique should adapt to the presence of motion. This chapter presents a motion adaptive de-interlacing technique incorporating a fuzzy motion detector and investigate how the quality of the results depends on the parameters of the technique. We also compare the technique numerically to other de-interlacing methods on several test-sequences.

Abstract: Here, we present two new neuron model architectures and one modified form of existing standard feedforward architecture (MSTD). Both the new models use self-scaling scaled conjugate gradient algorithm (SSCGA) and lambda–gamma (L–G) algorithm and entail the properties of basic as well as higher order neurons (i.e., multiplication and the aggregation functions). Of these two, compensatory neural network architecture (CNNA) requires relatively smaller number of inter-neuronal connections, cuts down on the computational budget by almost 50% and speeds up convergence, besides, gives better training and prediction accuracy. The second model sigma–pi–sigma (SPS) ensures faster convergence, better training and prediction accuracy. The third model (MSTD) performs much better than the standard feedforward architecture (STD). The effect of normalizing the outputs for training also studied here shows virtually no improvement, at low iteration level, say ∼500, with increasing range of scaling. Increasing the number of neurons beyond a point also shows to have little effect in the case of higher order neuron.The numerous simulation runs for the problem of satellite orbit determination and the complex XOR problems establishes the robustness of the proposed neuron models architectures.

Abstract: In many industrial applications of softcomputing, intelligent controls are important to accomplish high level tasks. Intelligent controls, however, need specific knowledge for each task. Therefore developing good memory is crucial to store the required knowledge efficiently and robustly. Neural network associative memories are the most suitable for the role because of their flexibility and content addressability. In this paper, first, we describe the basic concept of the neural network associative memories and the conventional learning algorithms. After pointing out some problems of the associative memories, we explain a novel learning algorithm, which is superior to the conventional ones. Finally, we introduce an associative memory suited for the intelligent controls and show the effectiveness by a number of computer simulations.

Abstract: This paper proposed a fast synthetic genetic algorithm (FSGA). The algorithm has faster convergence speed and higher computation precision, and the number of individuals and populations decreased, respectively. To overcome long training time of back propagation (BP) algorithm, this paper combined FSGA with BP, and applied the hybrid method to short-term economic dispatch of hydrothermal power system. The simulation results demonstrate that the hybrid method is effective and training time is short.

Abstract: A reinforcement learning-based neuro-fuzzy gait synthesizer, which is based on the GARIC (Generalized Approximate Reasoning for Intelligent Control) architecture, is proposed for the problem of biped dynamic balance. We modify the GARIC architecture to enable it to generate the trunk trajectory in both sagittal and frontal plane. The proposed gait synthesizer is trained by reinforcement learning that uses a multi-valued scalar signal to evaluate the degrees of failure or success for the biped locomotion by means of the ZMP (Zero Moment Point). It can form the initial dynamic balancing gait from linguistic rules, which are obtained from human intuitive balancing knowledge and biomechanics studies, and accumulate dynamic balancing knowledge through reinforcement learning, and thus constantly improve its gait during walking. The feasibility of the proposed method is verified through a 5-link biped robot simulation.

Abstract: The Choquet integral extended to negative functions can be defined in two ways, namley the asymmetric integral (usual Choquet integral) and the symmetric integral (also called the Šipoš integral). No such extension has been defined for the Sugeno integral. In this paper, after recalling the case of Choquet integral, we address the case of Sugeno integral, which we define in a purely ordinal framework.