Machine Learning Techniques for Smart City Applications: Trends and Solutions

Preface
Contents
1 Applying Deep Learning to Predict Civic Purpose Development: Within the Smart City Context
Abstract
1 Introduction
2 Methods
2.1 Dataset
2.2 Analyzed Variables
2.3 Deep Learning
2.4 Evaluation of Model Performance
2.5 Examining the Relationship Between Predictors and Predicted Outcomes
3 Experimental Results and Discussions
4 Concluding Remarks
Appendix: Supplementary Methods
References
2 Convolution Neural Network Scheme for Detection of Electricity Theft in Smart Grids
Abstract
1 Introduction
2 Background and Related Work
3 Model Implementation Details
3.1 Convolutional Neural Network
3.2 Batch Normalization
3.3 Max-Pooling and Flatten
3.4 Dropout
3.5 Dense Layer
3.6 Optimizers and Loss Function
3.7 Activation
4 Proposed Model Implementation
4.1 Data Attributes
4.2 Data Preprocessing
4.3 Model Design
5 Results
6 Discussion
7 Conclusion
References
3 Helping Hand: A GMM-Based Real-Time Assistive Device for Disabled Using Hand Gestures
Abstract
1 Introduction
2 Related Work
2.1 Hand Segmentation
2.2 Hand Feature Extraction
2.3 Hand Gesture Recognition
3 Proposed System
3.1 Gaussian Mixture Model (GMM)
3.1.1 Procedure
3.2 Hough Transform for Feature Extraction
3.2.1 Procedure
3.3 K-NN Classifier
4 Result and Discussion
5 Conclusion
References
4 A Review on Hand Gesture and Sign Language Techniques for Hearing Impaired Person
Abstract
1 Introduction
2 Findings on Hand Gesture and Sign Language Studies
2.1 Vision Based Study
2.2 Sensor Based Study
2.3 Hybrid Approaches
3 Discussion
3.1 Image Acquisition Devices
3.2 Performance Metrics
3.3 Sign Language Dataset
4 Hand Gesture and Sign Language Prospect in Smart Health Aspect
5 Conclusion
Acknowledgements
References
5 DriveSense: Adaptive System for Driving Behaviour Analysis and Ranking
Abstract
1 Introduction
2 Autonomous Vehicles, Smart Cities and DriveSense
3 DriveSense
4 Methodology
4.1 System Architecture
4.2 Machine-Learning Models and Algorithms
4.2.1 Decision Tree Classifier
4.2.2 Random Forest Classifier
4.2.3 Boosted Tree Classifier
4.2.4 Logistic Regression
4.2.5 SVM Classifier
5 Data Acquisition of Driving Behaviour
5.1 Use of Sensors to Determine Driving Behaviour Traits
5.2 Detecting Rash Driving Behaviour Using IMUs
6 Dataset Study
7 Machine-Learning Models
7.1 Machine-Learning Models—Virginia Dataset
7.1.1 SVM Classifier
7.2 Machine-Learning Models—Mendeley Dataset
7.2.1 Boosted Tree Classifier
7.2.2 Random Forest Classifier
7.2.3 Decision Tree
7.2.4 Logistic Classifier
8 Simulation
9 Metrics
10 Results
11 Conclusion
References
6 Classification and Tracking of Vehicles Using Videos Captured by Unmanned Aerial Vehicles
Abstract
1 Introduction
2 Related Work
2.1 Vehicle Detection Based on Motion Features
2.2 Vehicle Detection Based on Appearance Features
2.3 Vehicle Detection with Deep Learning
2.4 Vehicle Tracking
3 Metrics
3.1 Object Detection Metrics
3.2 Tracking Metrics
4 The UTUAV Urban Traffic Dataset
4.1 UTUAV-A Dataset
4.2 UTUAV-B Dataset
4.3 UTUAV-C Dataset
5 Methodology for Urban Vehicles Detection and Tracking
5.1 Light Vehicles
5.2 Motorcycles
5.3 Heavy Vehicles
6 Experiments
6.1 Detection and Classification
6.2 Vehicle Tracking
7 Discussion
8 Conclusion and Future Work
Acknowledgements
References
7 Tracking Everyone and Everything in Smart Cities with an ANN Driven Smart Antenna
Abstract
1 Introduction
2 Smart City and Intelligent Communications
3 5G/6G Systems and Other Communications Systems in Smart City with Research Challenges
3.1 5G/6G Systems for Smart Cities
3.2 Other Communications Networks
3.3 Research Challenges for Smart Cities
4 A Fast and Light ANN Enabled Antenna for Internet of Everything for a Smart City
4.1 ANN Antenna Model
4.2 Localization with ANN Enabled Antenna
4.3 Beamforming with SNWOM
4.4 Tracking with ANN Enabled Smart Antenna
4.5 Application of ANN Smart Antennas in Smart Cities
4.5.1 Smart Energy with ANN Enabled Smart Antennas
4.5.2 Smart Transportation with ANN Enabled Smart Antennas
4.5.3 Sensing Human Activity with ANN Enabled Smart Antennas
5 Summary
References
8 Wavelet-Based Saliency and Ensemble Classifier for Pedestrian Detection in Infrared Images
Abstract
1 Introduction
2 Proposed Methodology
2.1 Wavelet Decomposition and Feature Map Generation
2.2 Local Denary Pattern
2.3 Overlapping Pedestrian Detection
2.4 LogitBoost Classifier
2.5 Performance Measures
3 Results and Discussion
3.1 Wavelet Transform-Based Saliency Region Construction—Results
3.2 Pedestrian and Non-Pedestrian Classification Results
4 Conclusion
References
9 A Survey of Emerging Applications of Machine Learning in the Diagnosis and Management of Sleep Hygiene and Health in the Elderly Population
Abstract
1 Introduction
2 Etiology of Sleep
3 Types of Sleep Disorders
3.1 Insomnia
3.2 Sleep-Related Breathing Disorders (SRBD)
3.3 Central Disorders of Hypersomnolence
3.4 Circadian Rhythm Sleep–Wake Disorders
3.5 Parasomnias
3.6 Sleep-Related Movement Disorders
4 Aging-Related Sleep Disorders
4.1 Trends in Analysis of Sleep Disorders
4.1.1 Subjective Assessment of Sleep: Survey Questionnaire
4.1.2 Objective Assessment of Sleep: Polysomnography—The Gold Standard Test
4.1.3 Automated and Intelligent Machine Learning Systems for Sleep Assessment
4.1.4 Non-Contact and Unobtrusive Methods of Sleep Assessment in the Elderly
5 Conclusions and Future Scope
References
10 Smart City Traffic Patterns Prediction Using Machine Learning
Abstract
1 Introduction
2 Related Works
3 Methodology
3.1 Machine Learning Algorithms
3.1.1 Bagging (BAG)
3.1.2 K-Nearest Neighbor (KNN)
3.1.3 Multivariate Adaptive Regression Spline (MARS)
3.1.4 Bayesian Generalized Linear Model (BGLM)
3.1.5 Generalized Linear Model (GLM)
3.2 Performance Evaluation
3.3 Proposed Traffic Pattern Prediction System
4 Results and Discussion
5 Conclusion
References
11 Emergency Department Management Using Regression Models
Abstract
1 Introduction
2 Literature Survey
2.1 Types of Queuing Model Used
2.2 Hospital Beds’ Allotment
2.3 Outpatient Queuing
2.4 Patient Flow Management
2.5 External Factors
2.6 COVID-19 Pandemic and Its Impact in EDs
3 Methodology
4 Results and Discussion
5 Challenges
6 Conclusion
References
12 Machine Learning in Wind Energy: Generation to Supply
Abstract
1 Introduction
2 Need for Windfarms
3 Smart Cities and Wind Power
4 Methodology
5 Wind Forecasting
6 WindFarm Optimization
6.1 Optimization Using Random Search
6.2 Optimization Using Genetic Algorithm
6.3 Optimization Using Nelder-Mead and PSO
6.4 Novel Approach: Optimization Using Nelder-Mead and Genetic Algorithm
7 Fault Diagnosis in Transmission Line
8 Metrics
9 Results and Discussion
10 Conclusion
References
13 Multi-class Segmentation of Trash in Coastal Areas Using Encoder-Decoder Architecture
Abstract
1 Introduction
2 Proposed Methodology
3 Data
3.1 Study Area
3.2 UAV Details
3.3 Real Data
3.4 Synthetic Data
3.5 Randomly Generated Data (RGD)
4 Data Preprocessing
5 Deep Learning and Segmentation
6 Encoder-Decoder Architectures
6.1 ResNet—50 (Backbone)
6.2 U-Net
6.3 SegNet
7 Loss Function
7.1 Weighted Categorical Cross-Entropy
7.2 Dice Loss
7.3 Focal Tversky Loss
8 Evaluation Metrics
8.1 IoU
8.2 Dice Coefficient
9 Training
10 Results
11 Localization
12 Discussion
12.1 Architecture
12.2 Loss Functions
12.3 Segmentation Results
13 Summary
14 Conclusion
References
14 Learning Analytics for Smart Classroom System in a University Campus
Abstract
1 Introduction
2 Related Work
3 Proposed System
4 Learning Analytics in Smart Classroom
4.1 Application of ML Algorithms for Prediction
4.2 Techniques Used to Train the Model
5 Performance Evaluation and Model Selection
6 System Prototype
6.1 Hardware and Software Used
6.2 Implementation of Different Modules
7 System Evaluation
8 Conclusion and Future Works
References
15 Predictive Analytics for Smart Health Monitoring System in a University Campus
Abstract
1 Introduction
2 Related Work
3 Proposed System
4 Predictive Analytics in Smart Health
4.1 Applying Machine Learning Algorithms for Disease Prediction
4.2 Dataset Construction
5 Performance Evaluation of Each Model and Selection
5.1 Cold Flu Model
5.2 Hypertension Model
5.3 Diabetes Model
6 System Prototype
7 System Evaluation
8 Conclusion and Future Works
References
16 SysML-Based Design of Autonomous Multi-robot Cyber-Physical System Using Smart IoT Modules: A Case Study
Abstract
1 Introduction
2 Machine Learning-CPS-Related Research
3 System Description and Requirements
4 SysML Modeling of Autonomous Mobile Robots System
4.1 System Package Diagram
4.2 System Requirements Diagram
4.3 System Block Definition Diagram
4.4 System Internal Block Diagrams
4.5 System Activity Diagrams
4.6 System State Machine Diagrams
4.7 System Parametric Diagrams
5 Conclusions and Remarks
References
17 Vulnerabilities and Ethical Issues in Machine Learning for Smart City Applications
Abstract
1 Introduction
2 Vulnerabilities of ML in Smart City
2.1 Privacy
2.2 Data Security
2.3 Privacy and Public Administration
2.4 Scarcity of Skillful Professionals
2.5 Huge Capital
2.6 Unemployment
2.7 Economic Disparity
2.8 Artificial Stupidity
2.8.1 Racist Robots
2.8.2 Lack of Security
2.8.3 Robot Rights
3 Ethical Issues
3.1 Water Monitoring System
3.2 100 Smart Cities
3.3 Crime Prediction
3.4 Smart Grid
3.5 Occupancy Counting
3.6 GPS Tracking
3.7 Autonomous Transportation
3.8 Drone Applications
4 Discussions and Conclusion
References