A Simulation of Supervisory Human Operator Cognitive Processes (2002)

 

Project CTU0208813 supported by the Grant Agency of CTU, Prague, CR

 

Inicialization research project. A supervisory power plant operator model explores the physiological principles of influences of emotions and workload to cognition. Also user specific profile is considered with the aim of improvement real-time prediction of power plant operator behavior.The objectives of this project were the following: 1) to evaluate and validate quantitatively the safety assessment of power plant control under the safety critical conditions through simulation, 2) to monitor and predict human power plant operator behaviour under the normal and safety critical circumstances. Two basic simulation system architectures correspond to these two problems. Each problem leads to particular tasks, which can be analysed and studied through the simulation: analysis of personality type and user specific approaches to reasoning and task analysis, identifying of safety critical situations, which cannot be solved in a optimal way because of the limitation human operator’s resources; analysis of the human operator errors; studies of bi-directional interactions among human operator’s cognitive and emotional states on the one hand and the workload on the other hand.

One of the main goals of inicialization projects supported by the Grant Agency of CTU is the continuation of the inicialised research under supporting by other projects. This research continuus under the project financied from the Phare funds. See [MUSIP].

 

Project Staff:

 

Ing. Ladislava Janku*

Ing. Milan Šorf, Ph.D.

Ing. Marcela Fejtova

Ing. Martin Salac

Ing. Ladislava Dobias

Ing. Jan Macek

 

*Project manager

 

Project Overview:

 

1. Introduction

The exponential expansion and the tremendous growth of applying computational intelligence techniques in industrial simulations are apparent. Investigations on the simulation of cognition and man-machine interactions have recently received remarkable attention particularly for the following four application areas: design, safety assessment, training, and accident analysis.

Applications of models of cognitive behaviour in the area of design enable in essential to compare and evaluate the effectiveness of the different man-machine interfaces, evaluate their strengths and weaknesses. Cognitive approaches to the safety assessment require models of human machine interaction, simulation of human cognitive processes, simulation of plant and control systems and appropriate data and parameters reflecting the real world situation as well as possible. Applications of the cognitive models in training reflect the necessity to explore precious, structured human behaviour representations. Accident analysis is oriented to the identification of the causes of accident. There are two basic groups of accident causes - system failures and human operator errors. The analysis of the accidents caused by the human factor requires a special framework [1].

During the recent years, a wide spectrum of simulations of human behaviour and cognition, covering the whole area from artificial intelligence and fuzzy logic to control theoretic techniques, appeared on the scene [2].  However, there is still great deal the can be done in the area of a real-time monitoring of a human operator’s psychical stress and/or fatigue levels and integration of such estimators in simulation systems. This technique appears as a suitable approach promising control process fault minimization through the elimination of faults caused by the human factor. An experimental Yoshikawa and Takahashi’s study [3, 4] focused on a research in the line of human’s cognitive features, appears as a fundamental work in this field.  A concept of the two-way adaptive human machine interface was presented. Basic concept was later integrated in Mutual Adaptive Interface [5, 6, 7]. 

 

2. Different Approaches to the Simulation of Human Behaviour

In 1991, the model CAMEO (Cognitive and Action Modelling of Erring Operator, [8]) was presented which has been developed with the aim of simulating behaviour of a human operator in pretty complex working environments, such as nuclear power plants. This human operator behaviour model consisted of three sequentially connected modules - Perception and Recognition, Decision Making and Action modules, long term and short term memories and Attention Resource Controller, and was inspired by the human cognitive mechanism of the information processing and retrieval. It was implemented in G2 language [9].

Another working model was AIDE, which has been developed with the aim of simulating behaviour of military pilots [10]. It consisted of two basic elements: the model of competence and the model of performance. It has been developed using AI formalism.

The model CES (Cognitive Environmental Simulation, [11]) was focused on modelling cognitive behaviour of the nuclear power plant human operators under the emergency circumstances. This model is based on assumption that the people have limited resources, particularly when they are under the psychical stress. Not all the available knowledge can be utilised in an optimal way. The model CES consisted of the knowledge base representing all knowledge of the supervisory human operator and the basic processing mechanism, which is based on the use of analysts, which are AI operators equivalent to agents, already utilised by AIDE [12]. Three basic cognitive functions are simulated: monitoring the plant behaviour and its analysis (Behaviour Analysts), building explanation for unexpected conditions (Situation Analysts) and managing the response (Response Plan Analysts).

The system MIDAS (Man Machine Interaction Design and Analysis System, [13]) explores all the strengths of modular architectures.  The basic structure of this system contains the model of the controlled process and the model of human operator. These two models interact with each other dynamically. This architecture enables to simulate failures of the system under the control through the probabilistic modules.

Papenhuijzen [14, 15] presented the model of the navigator embedded in a generic architecture that discriminates between two major operator functions: between the function of the supervisor and the function of the actuator. Papenhuijzen’s model differentiated among three cognitive activities - state estimation, track planning and track following - represented by the specific sub-models. A link to control theory techniques can be observed in this approach.

 

3.  Analysis of the Human Behaviour Simulation System

This section refers to the process of designing a simulation system. We considered the simulation system development as a recursive process expanded into iterative loops.

3.1 Problem Formulation

The objectives of this study are the following: 1) to evaluate and validate quantitatively the safety assessment of power plant control under the safety critical conditions through simulation, 2) to monitor and predict human power plant operator behaviour under the normal and safety critical circumstances. Two basic simulation system architectures correspond to these two problems. Both are discussed in the section 3.5.

Each problem leads to particular tasks, which can be analysed and studied through the simulation: analysis of personality type and user specific approaches to reasoning and task analysis, identifying of safety critical situations, which cannot be solved in a optimal way because of the limitation human operator’s resources; analysis of the human operator errors; studies of bi-directional interactions among human operator’s cognitive and emotional states on the one hand and the workload on the other hand.

3.2 Description of the Human Behaviour Simulation System Architecture

As we said, there ate two main objectives of the simulation. The simulation system architecture shown on Fig. 1a reflects several aspects of the human machine interaction and can be utilised for the precious explorations of problems arising under the safety critical condition. On contrary, the second simulation architecture takes in account also the real-time monitoring of the power plant operator behaviour and a system ability to estimate current cognitive state of the operator (stress & fatigue levels).

Fig. 1a: A human behaviour simulation system architecture – human machine interaction simulation.

 

Fig. 1b: A human behaviour simulation system architecture – real-time monitoring and prediction of the human power plant operator behaviour.

 

4.  Analysis and Discussion of the Elements of the Human Behaviour Simulation System

This section describes the components of the simulation system in detail.  Both the component characteristics and the simulated interactions and dependencies are explored.

4.1 Intelligent User Profile

There are two important reasons for which the user profile technique should be utilised. The first one is implied by the appearance of the different personality types in the population of power plant operators. The second one reflects user specific approaches to problem solving, user specific preferences etc.  Most models described in section 2 have been developed on assumption, that the differences between human operators, influencing the processes of cognition, are inconsiderable. Such systems consider unique human operator models and cannot be used in tasks focused on a real-time prediction of behaviour of the particular human operator. We consider the ability of models to account specific aspects of behaviour, given by the particular personality types. From this point of view, each personality type is characterised by its own model of cognition.

 

Fig. 2: A structure of the Intelligent User Profile

 

The way in which the intelligent user profile influences the simulation process depends on the following interactions - the interaction among the intelligent user profile and cognitive processes or functions, the influence of the intelligent user profile to the emotion dynamics and the other interactions between the user profile and emotions, the influence of the environmental aspects to the weights of particular user profile items.

 

4.2 Real-Time Monitoring of Physiological Parameters

The presented simulation architecture takes in account estimated stress and/or fatigue levels of real power plant human operators. A vast amount of physiological data collected during the simulated experiments gives a satisfactory baseline for estimation procedures. Intelligent techniques have been applied to perform collected data analysis. All these techniques contain knowledge on relationships among physiological parameters (heart frequency, respiratory frequency, skin galvanic reaction, systolic and diastolic blood pressures, muscular tonus, etc.) and levels of stress or fatigue.

Because of impossibility to connect all the measurement equipment to one computer or to synchronise collected data sample per sample, a semi-distributed system for the data collection, based on the client-server paradigm, has been developed. Basic system structure is shown on Fig. 3.

Fig. 3: Semi-distributed architecture for data collection and analysis

 

The clients - data collection modules, collect data obtained from different measurement equipment. After the signal segmentation and time-stamps assignment to these segments, the simple or more sophisticated real-time pre-processing techniques are applied. Characteristic features are computed for each segment and, accompanied by the appropriate time-stamp, are sent to server. Simultaneously, the collected signal segments with time stamps are stored in the local databases. The way in which the data are processed at the level of clients - data collection modules - depends at any given time on the operations supported by the procedures implemented in these modules; they can be renovated in a very simple way. This semi-distributed approach to data collection and processing enables to make the best account of computational resources; it also reduces an amount of data transferred to the server. Server is also responsible for starting and stopping clients at any time and for the client synchronization. In section 5, the system implementation has been described in more detail.

 

4.3 Real-Time Estimation of the Stress and Fatique Levels

From the theoretical viewpoint, the techniques of real time estimation of the stress and fatigue levels can be considered as pattern classification methods. As we have already mentioned, all these techniques contain knowledge on relationships among physiological parameters and levels of stress or fatigue. Our previous research in essence involved application on artificial intelligence methods –  fuzzy rule system, feed-forward neural net, expert system [16]. We also verified the methodology of human operator’s psychical stress estimation based on measurement of physiological parameters and their processing through the application of computational intelligence approaches [17].

 

4.4 Simulation of Cognition

The basic constituents, which need to be included in each model of cognition, are cognitive functions and cognitive processes [2]. Cognitive functions are Perception, Interpretation, Planning and Execution. Cognitive processes are Memory/Knowledge Base and Allocation of Resources. Also the connections between cognitive functions and cognitive processes are considered. 

 

4.4.1 Short-term and Lon-term Memories

The short-term memory is modelled through the set of variables, which can be divided in several subsets: a subset cognitive representations of plant simulation user interface elements, a subset of cognitive representations of simulated environmental conditions and events, a subset of goals, a subset of performed actions, a subset of tasks to be solved, a   Examples: a variable reflecting the current state of the user interface element of the armature No. 002 of the primary cooling circuit; a current state of some user interface.

Long-term memory is modelled as a structure of organised knowledge. It specifies safety critical situation, standardised procedures, etc.

 

4.4.2 Perception

The mechanism of perception has been simulated through the cyclic scanning both of the variables of the user interface and of the variables of the simulated environmental conditions, elements and events. Contents of these variables are copied to the corresponding cognitive representation variables.  The model of attention has driven variable selection and timing of the scanning procedures.  

 

4.4.3 Observed Plant Behaviour Analysis

We consider a full ability of simulation to explain the perceived situation both in case of expected or unexpected findings. This cognitive function of interpretation has been simulated through the application of the diagnostic expert system.

 

4.4.4 Problem Solving Architecture and Plan Generation

A process of plan generation can leads to the identification of standardised procedures or, when these procedures are not available, to the complex planning process, in which the plan of actions, which have to be performed, is generated.  Each plan is generated with the respect to the identified state of a power plant simulation; a current plan, and the step of a current plan execution. These steps of execution take in account such situations in which the executed plan cannot be interrupted or modified.

 

4.4.5 Modelling Interaction among Cognition, User Profile and Emotions

This section deals with the basic principles of the simulation of influences of emotions and/or user specific characteristics to cognition.  There are three ways in which the cognition can be influenced by the user specific characteristic and emotions: knowledge management and activation, attention focusing, and limited resources.

 

4.4.6 Interactions between Workload and Cognition

This part of simulation has been inspired by the psychological principle that the human power plant operators have limited resources. They are not able to fully utilise all their knowledge in high workload or safety critical situations.

 

4.5 Simulation of Emotional Aspects

The influences of emotions of a real power plant operator to his/her cognitive processes and behaviour are not inconsiderable. The basic psychological principle is that the emotions can influence a quality of some cognitive functions and processes, e.g. a level of concentration, change the boundaries of the limited resources and influence the processes of knowledge management, e.g. some associations are reinforced, another are reduced. The seven essential emotions have been simulated.

An integration of emotions into the power plant operator model sets several new requirements to the whole simulation process. The way in which the emotions are activated depends on the following interactions - the interaction among emotions and cognitive processes or functions, the emotion dynamics, the interactions between user profile and emotions, and the interaction between emotions and environmental aspects, including simulated examples of power plant behaviour. Notice, that most of interactions are bi-directional.

The emotion dynamics simulation reflects essential biological neuro-hormonal interactions. The basic reference model structure was inspired through the discovered characteristics and functions of hypothalamus. This model is under the development.

 

4.7 Simulation of Plant and Control Processes

Observable power plant behaviour affects human operator cognitive activities. Essential power plant simulator architecture is shown on Fig. 4. Complexity and flexibility of the power plant simulation have been balanced to the complexity and flexibility of the human operator cognitive model.  In comparison to real power plant and control systems, the model embodies some substantial and significant simplifications.

 

 

Fig. 4: Simulator Architecture

 

This constituent of whole simulation system covers both the normal and safety critical examples. It consisted of a model of physical behaviour of the most important state variables and parameters, and of an interface supporting human machine interaction. Both plant and control process simulations are able to simulate dynamical behaviour of the real plant-control process system.

User interface is oriented to establish real connection with a real human operator or with the cognitive model of his or her. It provides information both through the graphical and command constituents. 

 

4.8 Simulation System Behaviour Monitoring

Except the human-machine interface discussed in the previous chapter, the simulation system embodies its own interface, which has been developed with the aim of the numerical and/or graphical presentation of the collected physiological data and the model of cognition behaviour. Though this interface forms an independent unit, it is not fully separated from the simulation process. However, it enables to interfere to the simulation process through the special commands. Whole interface is made up of the following constituents: collected data monitoring unit, collected data displaying unit, model of cognition monitoring unit, displaying of the selected characteristic of the model of cognition unit, displaying of the user profile parameters unit.

 

5. Implementation, Experiments and Results

This section describes several selected parts of the presented ongoing project, which have been already successfully implemented and tested.

 

5.1 Intelligent User Profile  - Computational Personality Type Classification

Goal: Human operator personality type classification.

Implementation: In contrast to the standard personality type classification process involving psychological investigation, presented approach is based on the human operator’s physiological parameters measurement (pulse frequency, systolic and diastolic blood pressures, skin resistance, muscular tonus). Personality type classification consists in neuroticism and tendency to risk behaviour classification. Five approaches have been applied and compared one to one – expert system (ES), machine learning (ML), combination of expert system and machine learning, fuzzy clustering, and fuzzy preference expert systems (PES) complemented by an aggregation mechanism [18, 19].

Results: Results acquired by individual methods do not differ very significantly. Table 1 gives an overview of the success rates of the applied AI methods.

 

Table 1: The success rate of the different classification approaches

 

Neuroticism

TRB

Neuroticism and TRB

Expert System

70%

68%

60%

Machine Learning

72%

72%

72%

Combination of ES and ML

94.9%

83%

81.4%

Fuzzy Clustering

60%

58%

55%

Fuzzy PES & aggregation

93%

90%

90%

 

These results are discussed in [18] in more detail.

 

5.2 User Specific Knowledge Modelling

Goal: Computational representation of user specific knowledge.

Implementation: This research activity is deeply involved with the psychology of cognition, particularly with the task analysis and solving problem areas. Simulation consisted of three essential components: modelling user specific task analysis procedures and approaches to problem solving, modelling user specific experience, and modelling additional static user specific knowledge. Notice, that this list is not exhaustive by any means. From the computational reasons, we reduce the count of the components forming user specific knowledge. All experiments assume simple examples of critical safety situation, which require exploring of power plant operator knowledge in an optimal way and a plan of actions generation. Modelling user specific task analysis procedures reflects the differences in cognition. Tested users have been demanded to solve selected examples of critical safety situations and to describe the cognitive process of a problem analysis and a plan of specific actions generation. User specific experience has been represented through the list of critical safety situations, applied successful and unsuccessful plans, and a counts and dates of the appearances of these situations. Last two items enable to simulate experience forgetting.  Additional user specific knowledge represents static knowledge, which is not covered in two previous knowledge simulation constituents, and which is not shared with the other power plant operators. 

Results: Made experiments have acknowledged an assumption of the improvement of a success rate and performance of the plan generation algorithms, when the user specific experience has been added to the simulation. The observed improvement was about 25% (average value) in performance measured in time needed for the plan generation. An influence of user specific cognitive approaches and processes can make the whole generation of the plan of actions best or worse. It’s apparent, that this technique enables make more precious prediction on of power plant supervisory operator behaviour. The obtained results have seemed as work of promise; but it has been already obvious, that the development of a methodology of the result evaluation is needed.

 

5.3 Real-Time Physiological Data Monitoring

Goal: Implementation of the data collecting system.

Implementation: The data collection system exploring the server-client architecture has been implemented in MATLAB. The basic server functionality covers starting and stopping the clients, their synchronization and storing collected data in a database. The basic client functionality includes physiological signal segmentation, assignment of the time-stamps to theses segments, computing features and characteristics from the measured data and sending these features to the server.  The measured signal is also stored locally – each client is able to store the time-stamped signal segments in a local database. Following modules are included: ECG Module, EOG Module, Skin Galvanic Response Module, Systolic Blood Pressure Module, Diastolic Blood Pressure Module, Muscular Tonus Module, Respiratory Frequency Module.

Results: This architecture enables to make the best of the parallel processing of the measured signals.  Numerical aspects of this implementation, the module structures, communication design, etc. are discussed in [20] in detail.  

 

5.4 Psychical Stress Level Estimation

Goal: Automatic human operator’s stress level estimation based on the real-time analysis of the collected physiological data.

Implementation: Two computational intelligence approaches were applied to estimate the psychical stress level – fuzzy rule system & feed-forward neural net. Both classification approaches performed classification to the categories described by the same three-value scale – low psychical stress, medium psychical stress, high psychical stress. Fuzzy rule system consisted of 27 fuzzy rules defined for 3 inputs & 3 outputs (classification to 3 classes). Applied neural net classifier consisted in 3-layer feed forward neural net  (3 neurons in input layer,  7 neurons in hidden layer & 3 neurons in output layer). Well-known back-propagation algorithm was used for this neural net learning. Two types of neuron output functions were used – linear & sigmoid functions [19]. A set of simple functions describing the relations among the psychical stress levels and simulated physiological parameters were created. These functions were derived from the real physiological data histories. Training data was generated using these functions, and it consisted of a matrix of 27 input vectors (blood pressures, heart rate, respiratory rate).

Results: Results obtained by both classifiers were very similar. Obtained results imply competency and correctness of this approach to actual psychical stress level estimation problem.

 

5.5 Interactions among Workload, Stress Level and Cognition

Goal: A simulation of influences of stress level and workload to cognitive processes

Implementation: Influences of workload and stress level have been simulated through the exploration of the following principles: resource limitation, knowledge activation, level of attention setting. While most human machine interaction systems assume unlimited and boundless capabilities of knowledge activation and a cognitive task performing, psychological studies advert to the fact that humans cannot explore or their resources and knowledge in an optimal way. This psychological principle is simulated through the limited resources. Only a specified amount of knowledge can be active at any given time. Only a specified amount of actions can be performed at any given time. The influence of attention has been simulated through the parallelism in thinking and through the association of such knowledge, which is not related to any solved problem.

Results: Made experiments are focused to the discovering of critical safety situations, which cannot be successfully solved by the power plant operators.

 

6. Conclusion

From the viewpoint of safety assessment in the area of supervision of large technological processes, the human-machine systems performing real-time power plant operator’s psychical state analysis and his/her future behaviour prediction as suitable systems for the safety risks minimization.

The objectives of this ongoing project have covered the evaluation of the safety assessment of power plant control under the safety critical conditions through simulation, and real-time monitoring and prediction of human power plant operator behaviour. The simulation system architectures have been presented and discussed in detail. The presented simulation system consisted of the following constituents: a simulation of plant and control processes, user interface model, a power plant operator model, a physiological data collection and evaluation system, a system performing estimation of the psychical and fatigue levels, prediction of power plant operator behaviour, and simulation system interface.

We described a supervisory power plant operator model, which explores the physiological principles of influences of personality types, emotions and workload to cognition. A concept of intelligent user profile has been presented. This concept has been implied by the appearance of the different personality types in the population and reflects also user specific characteristics, user specific approaches to problem solving, preferences etc. The theoretical foundations of model allow the formulation of rules describing interactions among emotions, user specific characteristics, workload, psychical stress and fatigue levels and cognition.

Because of this paper deals with the ongoing project, only particular experiments and results have been described in detail: computational personality classification techniques, modelling user specific knowledge, a system performing a real-time physiological data collection and processing, psychical stress level estimation techniques, modelling interactions among a workload, an estimated stress level and cognition. Made experiments seem to be very promising.

 

References

[1]       Senders, J.W, Moray, N.P: Human Error: Cause Prediction and Reduction. Lawrence Erlbaum Associates, N.J., U.S.A. 1991

[2]       Cacciabue, P.C.: Modelling and Simulation of Human Behaviour in System Control, Springer Verlag, 1998

[3]       Yoshikawa,H., Kitamura,M., Takahashi,M.: Development of Biocybernetic Analysis Systém (BIOCYS) for MMI - Related Cognitive Factor Analysis, Institute of Atomic Energy, Kyoto University Press, Japan, 1985

[4]       Yoshikawa,H., Takahashi,M., Arai,T.: Concept on Mutual Adaptive Interface and Related Experimental Study, Institute of Atomic Energy, Kyoto University Press, Japan, 1994

[5]       Takahashi,M., Kitamura, M., Yoshikawa,H.: Development of Real-Time Cognitive State Estimator, Institute of Atomic Energy, Kyoto University Press, Japan, 1995

[6]   Takahashi,M., Kubo,O., Yoshikawa,H.: Mutual Adaptive Interface: Basic Concept, Institute of Atomic Energy, Kyoto University Press, Japan, 1993

[7]   Kubo,O., Takahashi,M., Yoshikawa,H.: Mutual Adaptive Interface: Laboratory Experiments for Human State Estimation, Institute of Atomic Energy, Kyoto University Press, Japan 1993

[8]   Fujita, Y. et al.: Designing a Knowledge Based Operator Support System for Practical Applications, Nuclear Technology 95, 1991

[9]   Bar, A., Cohen, P.R., Feigenbaum, E.A: The Handbook of Artificial Intelligence – Volume IV., Addison-Wesley, Rading, MA, U.S.A, 1989

[10] Amalberti, R., Deblon, F.:Cognitive Modelling of fighter Aircraft Process Control; A Steps Towards an Intelligent Onboard Assistance System, International Journal of Man-Machine Studies, 36, 1992

[11] Woods, D.D., Roth, E.M., Pople, H.E.: Cognitive Environment Simulation: An Artificial Intelligence System for Human Performance Assessment. US-Nuclear Regulatory Commission, Technical Report NUREG-CR-4862. Washington DC 1987.

[12] Woods, D.D., Roth, E.M.: Symbolic AI Computer Simulations as Tools for Investigating Dynamics of joint Cognitive Systems, In Expertise and Technology: Cognition & Human-Computer Co-operation, Lawrence Erlbaum Associates, Hillsdale, N.J., U.S.A, 1995

[13] Corker, K.M., Smith, B.: An Architecture and Modelling for Cognitive Engineering Simulation Analysis: Application to Advanced Aviation Analysis, AAIA Conference on Computing in Aerospace, San Diego, CA, U.S.A. 1993

[14] Papenhuizen, R., Stassen, H.G.: On the modelling of the behaviour of a navigator. Proceedings of 8th Ship Control System Symposium, The Hague, The Netherlands, 1987

[15] Papenhuizen, R., Dijkhuis, T.: Evaluation of two human operator models of the navigator’s behavior. Proceedings 6th IFAC/IFIP/IFORS/IEA Symposium on Analysis, Design and Evaluation on Man-Machine Systems, Cambridge, MA, 1995.

[16] Sorf, M. - Eck, V. - Janku, L. - Lhotska, L.

Application of Neural Network for Stress Classification. In: The State of the Art in Computational Intelligence. Heidelberg : Physica-verlag, Germany, 2000

[17] Eck, V., Janku, L., Sorf, M., Novak, D., Fejtova, M., Roknic, J.: Applications of the Fuzzy Rule System and Neural Net to the Human Operator Stress Level Estimation. In: Proceedings. Zittau: Institut für Prozesstechnik und Messtechnik, Germany, 2001

[18] Janku, L., Sorf, M., Lhotska, L., Eck, V.: Five Approaches to Computational Personality Classification Problem, In: Kuncheva (Ed.), Proceedings of CIMA 2001, ICSC/NAISO conference, UK, 2001

[19] Janku, L., Lhotska, L.: Aggregation of Fuzzy Preferences as an Efficient Approach to the Computational Personality Type Classification Problem, in Mastorkais, N. (Ed.): Problems in Applied Mathematics and Computational Intelligence. Athens, World Scientific and Engineering Society Press, 2001

[20] Janku, L., Kremen, V.: Semi-Distributed Physiological Data Collection and Processing System, Research Report, Prague: CTU, Faculty of Electrical Engineering, Department of Cybernetics, 2002.

 

 

Publications:

-          Janků, L. - Lhotská, L - Eck, V. -  Matoušek, K. Some Aspects of the Simulation of Cognitive Behaviour of a Supervisory Human Operator in Complex Working Environment, inTrappl (ed.): Cybernetics and Systems, Austrian Society for Cybernetics Studies, Vienna, 2002, ISBN 3-85206-160-1

-          L. Janků  - V. Eck, L -  Lhotská - V. Křemen - J. Macek - L. Dobiáš - J. Dostál, P. Stejskal (EDOST s.r.o.): Analýza nehod a stanovení spolehlivosti řízení průmyslových systémů: aplikace simulací kognitivních procesů řídicího operátora; in Štěpánková, O. - Lhotská, L. - Krautwurmová, H. (editors): Intelligent Methods for Quality Improvement in Industrial Practice. Prague : CTU FEE, Department of Cybernetics, The Gerstner Laboratory, 2002. 143 p. ISSN 1213-3000.

 

This research was performed in co-operation with Dr. Ing. Lenka Lhotska and Dr. Ing. Vladimir Eck. The work of them was funded by the Ministry of Education of Czech Republic under the research program ‘Transdisciplinary Biomedical Engineering Research’, No.  MSM 210000012.