Intelligent and Safe Computer Systems in Control and Diagnostics

Preface
Acknowledgment
Contents
AI in Medicine
Trustworthy Applications of ML Algorithms in Medicine - Discussion and Preliminary Results for a Problem of Small Vessels Disease Diagnosis
1 Introduction
2 Diagnosis of Small Vessel Disease - Fundamentals
3 The Need for Trustworthiness of AI-Based Systems
4 ML-Based System Development
5 Small Vessels Disease Diagnosis - Preliminary Results
6 Concluding Remarks
References
Machine-Aided Detection of SARS-CoV-2 from Complete Blood Count
1 Introduction
2 Related Work
3 Our Solution
3.1 Data Collection
3.2 Data Preprocessing
3.3 Architectures
4 Experiments and Results
4.1 Experimental Setup
4.2 Baseline Training on UCC and Zenodo Datasets
4.3 Unbalanced vs Balanced Training
4.4 Impact of Joined Learning with an Additional Dataset
5 Discussion
6 Conclusion
References
Automatic Breath Analysis System Using Convolutional Neural Networks
1 Introduction
2 A Brief Overview of Similar Systems
3 Datasets
4 Breath Analysis System
5 Tests
6 Conclusions
References
Bridging Functional Model of Arterial Oxygen with Information of Venous Blood Gas: Validating Bioprocess Soft Sensor on Human Respiration
1 Introduction
1.1 Historic Context
1.2 Related Work
2 Methods
2.1 Clinical Study Conditions and Hardware
2.2 Model for Partial Pressures of Oxygen and Carbon Dioxide
3 Results
4 Conclusions
References
COVID-19 Severity Forecast Based on Machine Learning and Complete Blood Count Data
1 Introduction
2 Related Work
3 Our Solution
3.1 Data Collection
3.2 Data Preprocessing
3.3 Architectures
4 Experiments and Results
5 Discussion
6 Conclusion
References
Computer Diagnosis of Color Vision Deficiencies Using a Mobile Device
1 Classification of Color Vision Deficiencies
2 Computer Test for CVD
3 Solution
4 Conclusion and Future Work
References
Cybersecurity
Simulation Model and Scenarios for Testing Detectability of Cyberattacks in Industrial Control Systems
1 Introduction
2 Description of the Experimental Stand
3 Description of the Simulator
3.1 Overall Structure
3.2 Disturbances, Process and Cyber Faults Simulation
3.3 Possible Hardware in the Loop Configurations
4 Example of Cyber-Attack and Simulation of the System Performance
5 Conclusions
References
Functional Safety Management in Hazardous Process Installations Regarding the Role of Human Operators Interacting with the Control and Alarm Systems
1 Introduction
2 Defining Safety Functions for Reducing Risks
3 Layered Protection System in Hazardous Industrial Plants
4 Incorporating Cognitive Aspects in Human Reliability Analysis
4.1 Human Factors and Systems Cognitive Engineering
4.2 Human Behaviour Types
4.3 Including Cognitive Aspects in Human Reliability Analysis
4.4 Human Reliability Analysis in Context of Accident Scenarios
5 Case Study
5.1 Defining Accident Scenarios in Layered Protection System
5.2 Alarm System Design Issues to Meet Functional Safety Criteria in Context of Human Reliability Analysis
6 Conclusions
References
Controller Modelling as a Tool for Cyber-Attacks Detection
1 Introduction
2 Cyber-Attack Detection in Control Systems
3 Controller Modelling
3.1 Linear Model
3.2 Neural Network
3.3 Comparison of the Models
4 Case Study
5 Conclusions
References
Comparison of Traditional and Elliptic Curves Digital Signatures Providing the Same Security Level
1 Introduction
2 Digital Signature Algorithms
2.1 Schemes Based on Discrete Logarithm Complexity
3 Elliptic Curve Specific Signature Schemes
3.1 ElGamal Digital Signature Based on Elliptic Curves
4 Security Level of Signature Schemes
5 Experimental Comparison of Signature Schemes
5.1 Experiment Setup
5.2 Results
5.3 Results Analysis
6 Conclusions
References
Fundamental Concepts of Modeling Computer Security in Cyberphysical Systems
1 Introduction
2 Identifying the Attack Surface
2.1 Basic Terminology
2.2 Defining Security Services
3 Modeling Approach
3.1 An Overview
3.2 The NFR Approach
3.3 Simulation Modeling with Monterey Phoenix
3.4 Penetration Testing with Shodan Internet Search Engine
4 Integrating the Simulation and Pentesting into the NFR
4.1 Laboratory SCADA Equipment
4.2 Integration of the Simulation and Pentesting with the NFR
5 Conclusion
References
Artificial Neural Networks
Training of Deep Learning Models Using Synthetic Datasets
1 Introduction
2 Applied Methods and Techniques
2.1 Technical Details
2.2 Collecting 3D Models
2.3 Synthetic Dataset Generation
2.4 Validation Dataset Generation
2.5 Neural Network Architecture
2.6 Transfer Learning via Fine-Tuning
2.7 Scene Parameters Optimization
2.8 Network Parameters and Architecture Optimization
2.9 Validation
3 Results
3.1 The Role of Scene Organization in the Learning Process
3.2 Impact of Object Texture Properties on the Accuracy of a Neural Network
3.3 Impact of Camera Position on the Accuracy of a Neural Network
3.4 Network Architecture and Hyperparameters Optimization
3.5 PointRend Network
4 Conclusion
References
Autonomous Perception and Grasp Generation Based on Multiple 3D Sensors and Deep Learning
1 Introduction
2 Methods
2.1 Camera Setup
2.2 Camera Calibration
2.3 Point Cloud Merging
2.4 Converting a Point Cloud to an RGB Image
2.5 Instance Segmentation
2.6 Generating the Robotic Grips
2.7 Grasp Filtration Using GraspFilter
3 Results
3.1 Point Cloud Merging
3.2 OrthoView
3.3 Instance Segmentation
3.4 Initial-Grasps Generation
3.5 Initial-Grasps Filtration by GraspFilter
4 Discussion
References
Open-Set Speaker Identification Using Closed-Set Pretrained Embeddings
1 Introduction
2 Proposed Approach
2.1 Rationale
2.2 Data Flow
2.3 Speaker Recognition Backbone
3 Embedding-Based Classification of Speakers
4 Results
5 Summary
References
Condition-Based Monitoring of DC Motors Performed with Autoencoders
1 Introduction
2 Related Works
3 Overview
3.1 Autoencoders
4 Experiment Setup
4.1 Hardware and Software
4.2 Application
5 Results
5.1 Parameters
5.2 Datasets
5.3 Single Autoencoder, Single Work Point
5.4 Multiple Autoencoders, Multiple Work Points
5.5 Health Indicator and Signal Correlation
5.6 Comparison with Classical Methods
6 Conclusion
References
Estimation of Mass Flow Rates of Two-Phase Flow Using Convolutional Neural Networks
1 Introduction
2 Experimental Work
2.1 Experimental Setup
2.2 Methodology
3 Convolutional Neural Networks for Estimation
3.1 Image Classification
3.2 Data Augmentation
3.3 Training, Validation and Testing of the CNN
4 Results
5 Conclusions a Future Work
References
Recurrent Neural Network Based Adaptive Variable-Order Fractional PID Controller for Small Modular Reactor Thermal Power Control
1 Introduction
2 Problem Statement
2.1 Mathematical Model of SMR Nuclear Reactor
3 Research Method
3.1 Considered Controller Types
3.2 Adaptation Mechanism
4 Simulation Results
5 Conclusions
References
Fault Detection
LSTM Model-Based Fault Detection for Electric Vehicle's Battery Packs
1 Introduction
2 Research Methodology
2.1 Liquid Leakage and Liquid Intrusion Detection Method
2.2 Laboratory stand and experiment methodology
3 Results and Discussion
4 Conclusions
References
Remaining Useful Life Prediction of the Li-Ion Batteries
1 Introduction
2 RUL Prediction Methods
3 Fuzzy Logic Degradation Modeling Framework
4 Battery Remaining Useful Life Prediction
5 Validation of Remaining Useful Life Prediction
6 Conclusion Remarks
References
Detection of Multiple Leaks in Liquid Transmission Pipelines Using Static Flow Model
1 Introduction
2 General Characteristics of Diagnostic Methods
2.1 Method I
2.2 Method II
3 Experimental Data Acquired from the Laboratory Pipeline
3.1 Pipeline Stand
3.2 Conditions of Experiments
4 Results of Verification
4.1 Method I
4.2 Method II
5 Conclusion
References
Application of Bayesian Functional Gaussian Mixture Model Classifier for Cable Fault Isolation
1 Introduction
2 Bayesian Functional Gaussian Mixture Model
2.1 Spline Representation
2.2 Multiple Levels of Data in Diagnostics
2.3 Class Probability Reconstruction
3 Application to VSC DC Cable Diagnostics
3.1 Computational Setup
3.2 Example of Use
4 Sensitivity Analysis
5 Conclusions
References
Verification and Benchmarking in MPA Coprocessor Design Process
1 Introduction
2 Related Works
3 MPA Coprocessor
4 Design Process
5 Verification and Benchmarking Software
6 Conclusions
References
Sensor Fault Analysis of an Isolated Photovoltaic Generator
1 Introduction
2 Problem Statement
3 Modeling of the PVG
4 Proposed Diagnostic Approach and Results
4.1 PVG Around the Operating Point
4.2 Generation and Structuring
4.3 Analysis Through DCS Test
5 Results and Discussions
6 Conclusions
References
Systems Modeling
A Set-Based Uncertainty Quantification of Evolving Fuzzy Models for Data-Driven Prognostics
1 Introduction
2 Evolving Ellipsoidal Fuzzy Information Granules
2.1 Description
2.2 EEFIG-Based Degradation Modelling and RUL Estimation
3 Interval Set-Based Uncertainty Description
4 Case Study
4.1 Results and Discussion
5 Conclusions
References
Qualia: About Personal Emotions Representing Temporal Form of Impressions - Implementation Hypothesis and Application Example
1 Introduction
1.1 Qualia in Computational Models
1.2 The Contribution
2 Model of Human Emotions
3 Illustrative Simulation
3.1 The Influence of Sub-emotions on the Emotional State of the Agent (1st and 3rd Scenario)
3.2 Sub-emotion Creation (2nd Scenario)
4 Summary
References
Resistant to Correlated Noise and Outliers Discrete Identification of Continuous Non-linear Non-stationary Dynamic Objects
1 Introduction
2 Continuous-Time Modeling
2.1 Discrete-Time Approximation of Differential Equations
2.2 Non-linear Continuous-Time Models
3 Estimation Procedures
3.1 Least-Squares Method
3.2 Instrumental Variable Method
3.3 Least Absolute Values Method
4 Numerical Study
5 Conclusion
References
Neural Modelling of Dynamic Systems with Time Delays Based on an Adjusted NEAT Algorithm
1 Introduction
2 Problem Statement
3 dNEAT Algorithm
3.1 Initialisation
3.2 Crossover
3.3 Mutation
3.4 Fitness Function
4 Applications
4.1 Application 1
4.2 Application 2
5 Results
6 Conclusions
References
A Model-Based Approach for Testing Automotive Embedded Systems – A Preliminary Study
1 Introduction
2 Background and Related Works
2.1 Modelling Simulations for Embedded Software Development in the Automotive Industry
2.2 Embedded Software Testing as an Essential Safety, Quality and Reliability Phase
3 Research Methodology
3.1 Test Setup
3.2 Simulation Model
3.3 Research Approach
4 Results and Discussion
5 Conclusion
References
An Analysis of Observability and Detectability for Different Sets of Measured Outputs – CSTR Case Study
1 Introduction
2 Model of CSTR System
3 Analysis of observability and detectability
3.1 Results – CSTR Case Study 1
3.2 Results – CSTR Case Study 2
3.3 Results – CSTR Case Study 3
3.4 Results – CSTR Case Study 4
3.5 Results – CSTR Case Study 5
4 Conclusions
References
Adaptive, Robust and FTC Systems
The `Sense and Avoid' Aircraft System Based-on a Monocular Camera as the Last Chance to Prevent Accidents
1 Introduction
2 Measurable Image Parameters
3 Detectability and Avoidability
4 Multi Camera Systems
5 Projection Models
5.1 Disc Projection Model for Oblique Camera
5.2 Rectangle Projection Model for Oblique Camera
6 TTCPA and CPA Calculation in 2D and 3D
6.1 TTCPA and CPA in 3D
7 Extension of Method for Absolute Distance and Size and Application Guidelines for the Methods
8 Real Flight Results
9 Conclusion
References
Dynamic Positioning Capability Assessment for Ship Design Purposes
1 Introduction
1.1 Related Works
1.2 Motivation and Contribution
1.3 Structure of the Paper
2 Problem Definition
3 Methodology
3.1 Decision Variables
3.2 Constraints
3.3 Objective Function
3.4 Optimization Task
3.5 DP Capability Assessment
4 Results
4.1 Optimal Thrust Allocation
4.2 DP Capability Assessment
5 Conclusions
References
Degradation Tolerant Optimal Control Design for Linear Discrete-Times Systems
1 Introduction
2 Problem Formulation
3 Optimal Reconfiguration Control
3.1 Linear Quadratic Regulator
3.2 Linear Quadratic Tracker
4 EMA Application Example
4.1 Actuator Model
4.2 Model of Degradation
4.3 Results and Simulation
5 Conclusion and Future Work
References
.26em plus .1em minus .1emA Predictive Fault-Tolerant Tracking Control for Constrained Dynamic Systems
1 Introduction
2 Fault-Tolerant Tracking Controller Design
3 Simulation Results
4 Conclusions
References
A New Version of the On-Line Adaptive Non-standard Identification Procedure for Continuous-Time MISO Physical Processes
1 Introduction
2 Adaptive Model Identification Method
2.1 Modulating Functions Method
2.2 Re-identification Procedure for MISO Models
2.3 Exact State Observers
2.4 Adaptive Identification Algorithm
3 Experimental Results
4 Summary
References
Autonomous Systems Incidentally Controlled by a Remote Operator
1 Introduction
2 Types of Autonomy and Its Limitations
3 Virtual Teleportation
3.1 Passive vs Active VT
3.2 Subtasks to Implement VT
4 Autonomy Combined with Virtual Teleportation
4.1 Knowledge Base for an Autonomous System
4.2 Detection of Inability to Operate Autonomously
4.3 Learning from the Remote Operator
5 Example of the Application
6 Conclusions
References
Author Index