Positioning and Navigation Using Machine Learning Methods (Navigation: Science and Technology, 14)

Preface
Contents
1 GNSS Pseudorange Correction Using Machine Learning in Urban Areas
	1.1 Introduction
	1.2 Literature Review
	1.3 Methodology
		1.3.1 Several Common Inputs
		1.3.2 Labelling Process
	1.4 Static Experiments
		1.4.1 Algorithm Framework
		1.4.2 Data Collection
		1.4.3 Training Process of GBDT
		1.4.4 Position Calculation with Two Variations
		1.4.5 Static Case with Wide Road and High-Rise Buildings on One Side
	1.5 Dynamic Experiments
		1.5.1 Algorithm Framework
		1.5.2 Training Data Generation
		1.5.3 Random Forest (RF) Based Pseudorange Error Prediction
		1.5.4 Positioning with Two Variations
		1.5.5 Field Test and Analysis of Results
		1.5.6 PBC Analysis
		1.5.7 GBC Analysis
		1.5.8 Positioning Performance Evaluation
	1.6 Conclusion
	References
2 Deep Learning-Enabled Fusion to Bridge GPS Outages for INS/GPS Integrated Navigation
	2.1 Introduction
	2.2 INS/GPS Integrated Navigation System
		2.2.1 Inertial Navigation Coordinate Systems
		2.2.2 Kalman Filter Model
	2.3 The INS/GPS Fusion Algorithm Based on Kalman Filter
		2.3.1 INS Position Estimation Algorithm
		2.3.2 Differential Equations of INS Errors
		2.3.3 Design of Loosely Coupled INS/GPS Integrated Navigation System
	2.4 The INS/GPS Integrated Navigation Algorithm in GPS Denied Environments
		2.4.1 GRU Neural Network
		2.4.2 A Deep Learning-Assisted INS/GPS Integrated Navigation Structure
		2.4.3 The GI-NN Architecture
	2.5 Experimental Results and Analysis
		2.5.1 Simulation Experiment
		2.5.2 Field Tests
		2.5.3 Dataset Tests
	2.6 Conclusion
	References
3 Integrity Monitoring for GNSS Precision Positioning
	3.1 Introduction
	3.2 Developments of Integrity Monitoring
		3.2.1 Overview of Integrity Monitoring for Aviation
		3.2.2 Overview of Integrity Monitoring for Precise Positioning
	3.3 Integrity Monitoring Challenges for High-Precision Positioning
		3.3.1 FDE Algorithms
		3.3.2 User-End PL Calculation
	3.4 Preliminary Studies on Generalization of Integrity Monitoring Algorithms
		3.4.1 The DIA-MAP Method for FDE Purpose
		3.4.2 Stochastic Modelling for the Residual Tropospheric Delays
	3.5 Concluding Remarks
	References
4 Machine Learning-Aided Tropospheric Delay Modeling over China
	4.1 Introduction
	4.2 Data Description
	4.3 Methodology
		4.3.1 ZTD Calculation Using Real-Time GNSS and GFS Forecasts
		4.3.2 Calibrating GFS Forecasts Using Real-Time GNSS Dataset
		4.3.3 Polynomial Approximation
	4.4 Results and Discussion
		4.4.1 Evaluation of the GFS-Forecasted ZTDs
		4.4.2 Performance of the WAPTCs Under Challenging Conditions
	4.5 Conclusion
	References
5 Deep Learning Based GNSS Time Series Prediction in Presence of Color Noise
	5.1 Introduction
	5.2 Principle and Method
		5.2.1 Variational Modal Decomposition （VMD）
		5.2.2 Long Short Term Memory （LSTM）
		5.2.3 Dual Variational Mode Decomposition Long-Short Term Memory Network Model （DVMD-LSTM）
		5.2.4 Precision Evaluation Index
	5.3 Data and Experiments
		5.3.1 Data Sources
		5.3.2 Data Preprocessing
		5.3.3 VMD Parameter Discussion
	5.4 Experimental Results and Analysis
		5.4.1 DVMD-LSTM Prediction Results Analysis
		5.4.2 DVMD-LSTM Model Prediction Results and Precision Analysis
		5.4.3 Optimal Noise Model Research
	5.5 Conclusion
	References
6 Autonomous UAV Outdoors Navigation—A Machine-Learning Perspective
	6.1 Introduction
	6.2 Approaches of UAV Outdoor Navigation
		6.2.1 UAV Navigation Techniques: Technological Perspective
		6.2.2 UAV Navigation Techniques: Remarks
	6.3 ML-Based Autonomous UAV Navigation Techniques
		6.3.1 Operational Requirements
		6.3.2 The Operating Environment
		6.3.3 Cellular Navigation Problem Formulation
		6.3.4 Cellular ML-Based UAV Navigation
	6.4 Experimental Validation
		6.4.1 Field Test Experimentation
		6.4.2 Simulation Results and Analysis
		6.4.3 Overall Results and Analysis
		6.4.4 Discussing the Experimental Versus Simulation Results
	6.5 Conclusions
	References
7 Magnetic Positioning Based on Evolutionary Algorithms
	7.1 Introduction
	7.2 Magnetic Signal Measurement and Processing Methods
		7.2.1 Influence Factors Analysis of Magnetic Field Data
		7.2.2 Magnetic Data Processing Methods
	7.3 SPMP Based on an Enhanced Mind Evolutionary Algorithm
		7.3.1 Enhanced Mind Evolutionary Algorithm
		7.3.2 Definition of the Localization Model
		7.3.3 SPMP Experimental Results
	7.4 SBMP Using an Enhanced Genetic Algorithm-Based Extreme Learning Machine
		7.4.1 Extreme Learning Machine
		7.4.2 Enhanced Genetic Algorithm
		7.4.3 Localization Model Definition Using EGA-Based ELM
		7.4.4 SBMP Experimental Results
	7.5 Summary
	References
8 Indoor Acoustic Localization
	8.1 Introduction
	8.2 Robust Acoustic Signal Ranging
	8.3 Node Auto Calibration
		8.3.1 Other Method for Node Auto Calibration
	8.4 Acoustic Indoor Localization Algorithm
	8.5 Experimental Results
	8.6 Conclusion
	References
9 Scalable and Accurate Floor Identification via Crowdsourcing and Deep Learning
	9.1 Introduction
	9.2 Related Work
	9.3 Proposed Method
		9.3.1 Ground Floor Detection
		9.3.2 Automatic Fingerprint-Floor Level Association Via Crowdsourcing
		9.3.3 Floor Identification
	9.4 Experiments and Results
		9.4.1 Accuracy of IO Switch Detection
		9.4.2 Floor Identification Performance
		9.4.3 Effect of Different Smartphones
		9.4.4 Computational Cost
		9.4.5 Comparison with State-of-the-Art Methods
	9.5 Conclusion
	References
10 Indoor Floor Detection and Localization Based on Deep Learning and Particle Filter
	10.1 Introduction
	10.2 Related Works
		10.2.1 Localization Based on PDR
		10.2.2 Floor Detection
		10.2.3 DL in Localization Systems
	10.3 Proposed DL-Based Floor Detection
		10.3.1 User Activity Analysis and Floor Detection Scenario
		10.3.2 DL Model Selection, Data Processing, and Training Results
		10.3.3 Floor Decision Algorithm with Relative Pressure Map
	10.4 PDR-PF with Clustering
		10.4.1 Smartphone Sensor-Based PDR
		10.4.2 PF with Clustering and Correction
	10.5 Experiment Results
		10.5.1 DL-Based Floor Detection
		10.5.2 Multi-floor Indoor Localization
	10.6 Conclusion and Discussion
	References
11 An Indoor Wi-Fi Localization Algorithm Using BP Neural Network
	11.1 Introduction
	11.2 Wi-Fi Localization
		11.2.1 Fingerprint-Based Wi-Fi Localization
		11.2.2 Ranging-Based Wi-Fi Localization
		11.2.3 Angle-Based Wi-Fi Localization
		11.2.4 Machine Learning-Based Wi-Fi Localization
	11.3 Translation and Scaling of RSSI Vector
	11.4 Constructing the Ranging Model
	11.5 Position Determination
	11.6 Experimental Analyses
		11.6.1 Performance of RSSI Vector Translation and Scaling
		11.6.2 Performance of the Ranging Model
		11.6.3 The Effect of Distance Threshold
		11.6.4 Analysis of the Positioning Accuracy
	11.7 Conclusion
	References
12 Intelligent Indoor Positioning Based on Wireless Signals
	12.1 Introduction
	12.2 Challenges of Indoor Positioning
	12.3 Indoor Positioning Based on Machine Learning
		12.3.1 Supervised Learning
		12.3.2 Unsupervised Learning
		12.3.3 Semi-Supervised Learning
	12.4 Sensor-Fusion Positioning Techniques
		12.4.1 Sensor Fusion via Bayesian Filters
		12.4.2 Indoor Positioning via Trajectory Fusion
	12.5 Experimental Results and Performance Comparison
		12.5.1 Traditional Machine Learning
		12.5.2 Deep Learning
		12.5.3 Crowdsourcing Solutions
	12.6 Conclusion
	References
13 High Precision Positioning Algorithms Based on Improved Sparse Bayesian Learning in MmWave MIMO Systems
	13.1 Introduction
	13.2 System Model
	13.3 Conventional Parameter Estimation Algorithms
		13.3.1 Subspace Algorithms, MUSIC and ESPRIT
		13.3.2 Iterative Algorithm
		13.3.3 Statistical Sparse Recovery
		13.3.4 Discussion
	13.4 Improved Sparse Bayesian Learning Algorithm
		13.4.1 Sparse Bayesian Learning Formulation
		13.4.2 Grid Refinement
		13.4.3 SBL with Grid Refinement
		13.4.4 Comparison of Algorithm Complexity
	13.5 Numerical Result
		13.5.1 Verification of the Adaptive Grid Refinement Method
		13.5.2 Comparison of Positioning Results
	13.6 The Result of Measured Data
	13.7 Summary
	References
14 UWB Non-line-of-Sight Propagation Identification and Localization
	14.1 Introduction
	14.2 NLOS Identification Methods Based on Machine Learning
		14.2.1 Method Description
		14.2.2 Experimental Analysis
	14.3 Localization Based on Distance Error Mitigation
		14.3.1 Modeling of Ranging Error
		14.3.2 Improved Chan-Kalman Localization Algorithm
	14.4 Conclusions
	References