Big data analytics for cyber-physical systems : machine learning for the internet of things
Big Data Analytics in Cyber-Physical Systems: Machine Learning for the Internet of Things examines sensor signal processing, IoT gateways, optimization and decision-making, intelligent mobility, and implementation of machine learning algorithms in embedded systems. This book focuses on the interacti...
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | E-Book Buch |
| Sprache: | Englisch |
| Veröffentlicht: |
Amsterdam
Elsevier
2019
|
| Ausgabe: | 1 |
| Schlagworte: | |
| ISBN: | 9780128166376, 0128166371 |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Inhaltsangabe:
- 6.1. Preprocessing and inference chain -- 6.2. Data selection and evaluation approach -- 7. Evaluation and results -- 7.1. Parameter tuning with cross-validation (zone 1) -- 7.2. Performance evaluation (zone 2) -- 7.3. Discussion -- 8. Conclusion and future work -- References -- Chapter 6: Portable implementations for heterogeneous hardware platforms in autonomous driving systems -- 1. Advanced driver-assistance systems -- 2. Programming challenges -- 2.1. Understanding scaling-behavioral details -- 2.1.1. Metrics -- 2.1.2. Case study -- 2.1.3. Characteristics -- 2.2. Migration and architecture mapping -- 3. Parallel programming approaches -- 3.1. Basic multicore parallelization -- 3.1.1. SIMD vectorization -- 3.1.2. Automated parallelization tools -- 3.2. Programming heterogeneous architectures -- 3.2.1. Programming accelerators -- 3.2.2. Architecture mapping -- 4. Unification -- 4.1. Unifying heterogeneous software-code -- 4.2. Coherent shared memory -- 4.3. Abstracting middleware layers -- 5. Summary -- References -- Chapter 7: AI-based sensor platforms for the IoT in smart cities -- 1. Introduction -- 2. Function units of an IoT sensor -- 3. More than one sensor element -- 4. The communication interface -- 5. Embedded O/S requirements -- 6. Artificial intelligence embedded -- 7. Classification and regression using machine learning algorithms -- 8. Learning process required -- 9. AI-based IoT sensor system -- 10. Decentralized intelligence -- 11. Conclusions -- References -- Chapter 8: Predicting energy consumption using machine learning -- 1. Introduction -- 2. Data profiling -- 2.1. Data integration -- 2.2. Outlier detection -- 2.3. Feature selection -- 3. Learning from data -- 3.1. Multiple regression -- 3.2. Artificial neural networks -- 3.3. Valuation metrics -- 4. Related work -- 4.1. Data preprocessing -- 4.2. Procedure models
- Front Cover -- Big Data Analytics forCyber-Physical Systems: Machine Learning for the Internet of Things -- Copyright -- Contents -- Contributors -- Foreword -- Acknowledgments -- Introduction -- Chapter 1: Data analytics and processing platforms in CPS -- 1. Open source versus proprietary software -- 2. Data types -- 3. Easy data visualization using code -- 4. Statistical measurements in CPS data -- 5. Statistical methods, models, and techniques: Brief introduction -- 6. Analytics and statistics versus ML techniques -- 7. Data charts -- 8. Machine logs analysis and dashboarding -- 9. Conclusion -- References -- Chapter 2: Fundamentals of data analysis and statistics -- 1. Introduction -- 2. Useful software tools -- 2.1. Software for statistical computations -- 2.2. Software for interactive graphical data representation -- 3. Fundamentals of statistics -- 3.1. Features and scalings -- 3.1.1. Nominal, ordinal, and metric features -- 3.1.2. Discrete and continuous features -- 3.2. Characterization of univariate distributions -- 3.2.1. Nominal features -- 3.2.2. Ordinal features and quantiles -- 3.2.3. Metric features-Grapical representation -- 3.2.4. Metric features-Characteristic numbers -- 3.3. Characterization of multivariate distributions -- 4. Regression: Fitting functional models to the data -- 4.1. Linear regression -- 4.1.1. Linear least-squares regression -- 4.1.2. Linear total least-squares regression -- 4.2. Nonlinear regression -- 4.2.1. Polynomials -- 4.2.2. Population growth-Logistic function -- 4.3. Logistic regression -- 5. Minimizing redundancy: Factor analysis and principle component analysis -- 6. Explore unknown data: Cluster analysis -- 6.1. k-Means -- 6.2. Hierarchical clustering -- 6.3. Density-based clustering -- 6.4. Outlier detection -- 7. Conclusion -- References
- 5.4. Results -- 6. Conclusion -- References -- Chapter 11: Machine learning-based artificial nose on a low-cost IoT-hardware -- 1. Introduction -- 2. Related work -- 3. Temperature-modulated gas sensing -- 3.1. Principle component analysis -- 4. Support vector machine -- 4.1. Separable case -- 4.2. Nonseparable case -- 4.3. ECOC model -- 5. Feature selection -- 5.1. Wrappers -- 6. Technical implementation -- 6.1. Equipment -- 6.1.1. Hardware IoT device -- 6.1.2. Semiconductor gas sensor BME680 -- 6.1.3. Cloud computer -- 6.2. Measurement process -- 7. Experimental setup and results -- 8. Conclusion -- References -- Glossary -- Chapter 12: Machine Learning in future intensive care-Classification of stochastic Petri Nets via continuous-time Markov ... -- 1. Introduction -- 1.1. Digitalization of the intensive care unit -- 1.2. Petri nets and data analytics for clinical data -- 1.3. Our contribution -- 2. Background -- 2.1. Stochastic Petri nets -- 2.2. Continuous-time Markov chains -- 3. Methodology -- 3.1. Quantization -- 3.2. EM-algorithm for CTMC parameter learning -- 3.3. SPN topology estimation -- 3.4. Support vector machine -- 4. Simulation setup -- 4.1. SPN structure -- 4.2. Training procedure -- 5. Results -- 6. Conclusion and future work -- References -- Chapter 13: Privacy issues in smart cities: Insights into citizens perspectives toward safe mobility in urban environments -- 1. Introduction -- 2. Automated driving and privacy issues -- 2.1. Willingness to share data in automated driving, data types, and storage -- 2.2. Acceptance of cameras on public traffic routes -- 3. Crime surveillance and privacy issues -- 3.1. Acceptance of surveillance technologies at different locations -- 3.2. Trade-off between safety and privacy regarding acceptance of surveillance technologies -- 4. Conclusions, outlook, and future research directions
- Chapter 3: Density-based clustering techniques for object detection and peak segmentation in expanding data fields -- 1. Introduction -- 2. Related work -- 3. A brief introduction to density-based clustering -- 4. Formal extensions of density-based clustering -- 4.1. Operation on lattice-based data -- 5. Clustering strategy for time-expandable data sets -- 5.1. Legacy expansion approach -- 5.2. Horizons of processing areas with incomplete data -- 5.3. Algorithmic strategy for data with expansion -- 6. Evaluation and results -- 7. Conclusion -- References -- Chapter 4: Security for a regional network platform in IoT -- 1. Introduction -- 2. Regional network security -- 2.1. Regional network -- 2.2. Threats and design requirements -- 3. Proactive distributed authentication framework for a regional network -- 3.1. Automatic trustworthy registration -- 3.2. Distributed access authentication -- 3.3. Mutual authentication procedure -- 3.4. Temporary registration and authentication in the event of a disaster -- 3.5. Handover authentication procedure -- 3.6. Blacklist sharing procedure -- 4. Discussion -- 5. Function implementations -- 6. Network setup and performance evaluations -- 6.1. Network setup -- 6.2. Metrics and experimental setup -- 6.3. Experimental results -- 7. Conclusions -- References -- Further reading -- Chapter 5: Inference techniques for ultrasonic parking lot occupancy sensing based on smart city infrastructure -- 1. Introduction -- 2. Related work -- 3. Fundamentals and background -- 3.1. Introduction to support vector machines -- 3.2. Introduction to principal component analysis -- 4. System setup and architecture -- 4.1. Platform and embedded system concept -- 4.2. Ultrasonic sensing properties and synchronization -- 4.3. Real-world evaluation scenario -- 5. Data annotation and trainging methodoloy -- 6. Proposed method
- References
- 4.3. Prognosis -- 4.4. Evaluation -- 4.5. Conclusion -- 5. Further thoughts -- Acknowledgments -- References -- Chapter 9: Reinforcement learning and deep neural network for autonomous driving -- 1. Introduction -- 1.1. Motivation -- 1.2. Objective -- 1.3. Related work -- 2. Signal model -- 2.1. Scenario -- 2.2. Sensor signal model -- 3. Machine learning -- 3.1. Deep learning and neural networks -- 3.1.1. Feed-forward network -- 3.1.2. Recurrent neural network -- 3.1.3. Loss function -- 3.2. Reinforcement learning -- 3.2.1. Markov decision process -- 3.2.2. Reward -- 3.3. Policy -- 3.3.1. Value functions -- 3.3.2. Temporal difference learning -- 3.3.3. Policy gradient method -- 3.3.4. Actor-critic algorithm -- 3.3.5. On-policy and off-policy -- 3.3.6. Exploration vs exploitation -- 3.4. Deep deterministic policy gradients -- 3.4.1. Experience replay -- 3.4.2. Separate target network -- 3.4.3. Update rule for weights of the neural network -- 4. Simulation -- 4.1. Developing a scenario -- 4.2. Dynamic environment -- 4.3. Testbed -- 5. Conclusion and future work -- References -- Chapter 10: On the use of evolutionary algorithms for localization and mapping: Infrastructure monitoring in smart cities ... -- 1. Introduction -- 1.1. Monitoring infrastructure -- 1.2. State-of-the-art -- 1.3. Chapter organization -- 2. Centralized offline localization and mapping -- 3. Evolutionary localization and mapping -- 3.1. EVOLAM overview -- 3.2. MOEA -- 3.3. Simulation -- 3.4. Localization -- 3.4.1. Measurement model -- 3.4.2. Motion model -- 3.4.3. Maximum a posteriori optimal localization -- 3.5. Environment mapping -- 4. Multiobjective evolutionary algorithms -- 4.1. Dominance-based MOEA -- 4.2. Indicator-based MOEA -- 4.3. Other approaches -- 5. Case-study -- 5.1. Problem statement -- 5.2. Objective functions -- 5.3. Adopted MOEA -- 5.3.1. Simulation