This thesis is structured in four parts for a total of ten chapters. The first part, introduction and review (Chapters 1 to 4), presents an extensive state-of-the-art review of both symbolic and connectionist GI methods, that serves also to state most of the basic material needed to describe later the contributions of the thesis. These contributions constitute the contents of the rest of parts (Chapters 5 to 10). The second part, contributions on symbolic and connectionist techniques for regular grammatical inference (Chapters 5 to 7), describes the contributions related to the theory and methods for regular GI, which include other lateral subjects such as the representation oí. finite-state machines (FSMs) in recurrent neural networks (RNNs).The third part of the thesis, augmented regular expressions and their inductive inference, comprises Chapters 8 and 9. The augmented regular expressions (or AREs) are defined and proposed as a new representation for a subclass of CSLs that does not contain all the context-free languages but a large class of languages capable of describing patterns with symmetries and other (context-sensitive) structures of interest in pattern recognition problems.The fourth part of the thesis just includes Chapter 10: conclusions and future research. Chapter 10 summarizes the main results obtained and points out the lines of further research that should be followed both to deepen in some of the theoretical aspects raised and to facilitate the application of the developed GI tools to real-world problems in the area of computer vision.

En la darrera dècada, el disseny de sistemes de supervisió per a processos industrials ha rebut força atenció. Aquest fet s'explica per l'augment de la demanda en prestacions, flexibilitat i seguretat causada per la creixent conscienciació en qüestions de qualitat , legislació ambiental, i productivitat. L'interès principal d'aquesta Tesi es centra en el disseny d'un Sistema de Supervisió Intel·ligent (SIS). La tasca encomanada al SIS és la vigilància del procés, que consisteix en identificar i notificar a l'operador el seu estat de funcionament. Així per exemple, si el SIS detecta un estat de fallada, l'operador podrà emprendre les accions correctores adequades.La originalitat d'aquest treball es basa en la proposta d'una metodologia de disseny basada en la interpretació d'esdeveniments significatius, detectats en els senyals mesurats. En aquesta metodologia, l'expert té un paper rellevant, proporcionant el seu coneixement heurístic del procés. Es proposa una arquitectura de supervisió estructurada en dos nivells: la interfície i el supervisor.La interfície és l'encarregada d'abstraure informació significativa del procés, a partir de l'anàlisi dels senyals mesurats. Per aquest nivell, s'ha formalitzat una metodologia de disseny fonamentada en el paradigma de les finestres lliscants. Una finestra és un conjunt de dades consecutives d'un senyal, caracteritzada per una amplada i un desplaçament. Aplicant un càlcul determinat sobre aquestes dades és possible obtenir una nova dada. Així, l'anàlisi de diverses finestres consecutives proporciona un nou senyal sobre el que es pot aplicar un procediment similar. Continuant aquest procés iteratiu, es pot arribar a abstraure un senyal amb un contingut prou significatiu per a l'expert. Aquesta informació és transmesa al supervisor en forma d'esdeveniments.El supervisor s'encarrega d'interpretar els diversos encadenaments d'esdeveniments observats, i notifica l'estat de funcionament del procés a l'operador. Per aquest nivell, s'ha desenvolupat una metodologia per a l'obtenció d'un model comportamental de la planta, com a màquina d'estats finits. El procediment de modelatge proposat considera la definició de models associats a diversos components del procés i a les interaccions que s'estableixen entre ells. L'autòmat final s'obté per composició de tots aquests models.Ambdós nivells es nodreixen del coneixement heurístic de l'operador: definint els esdeveniments significatius a detectar mitjançant l'anàlisi dels senyals, i associant els encadenaments esperats a estats de funcionament.La metodologia de disseny del SIS ha estat validada satisfactòriament mitjançant l'aplicació a un procés biotecnològic i a una estació de mecanitzat.Diversos resultats d'aquest treball han estat presentats a congressos nacionals i internacionals. / OF THE THESISOver the last decade, the design of supervisory systems for industrial processes has received a great deal of attention. One reason for this is the increased demands on performance, flexibility, and safety caused by increased awareness, environmental regulations, and customer-driven productivity.The main purpose of this Thesis is the design of a Supervisory Intelligent System (SIS). The proposed SIS is concerned with process monitoring, consisting in process functional state identification and its notification to the operator. For example, on faulty state detection, the operator could apply the appropriate corrective action.The originality of this work consists in proposing a design methodology based on the interpretation of significative events, detected from measured signals. In this methodology, heuristic knowledge supplied by operators and experts is very important. A two level supervision architecture is proposed: the interface and the supervisor.The interface deals with abstracting significative process information by means of measured signals data analysis. For this level, a design methodology based on the sliding window paradigm has been formalised. A window is a signal data set, characterised by a duration and a sliding time. Running a specific computation on this data set produces new data. So, the analysis of different consecutive windows leads to a new signal on which a similar procedure can then be applied. Following this iterative process, a signal can be derived which carries enough significative information according to the operator. This information is transmitted to the supervisor as events. The supervisor translates observed event sequences into process states, which later are notified to the operator. For this level, we propose a design methodology that supplies a behavioural plant model by means of a finite state machine description. This procedure is based on modelling process components and interactions between them. The final automaton is obtained by the composition of all these models.Both SIS levels depend on the available operator heuristic knowledge: defining significative events which are to be detected using data analysis, and linking expected event sequences to functional states.The SIS design methodology has been successfully validated through its application to a biotechnological process and a milling machine.Several results in this dissertation have been presented at national and international conferences.