Wireless Localization Techniques

Preface
Introduction
Contents
1 Why and What?
1.1 Wireless Localization
1.1.1 Why Not GPS?
1.1.2 What Is Wireless Localization?
1.1.3 Issues to Be Studied in the Book
1.2 Background and Literature Review
1.2.1 Early Localization System
1.2.2 Mobile Crowdsensing and Wi-Fi Fingerprints
1.2.3 The Accuracy of RSSI Localization
1.2.4 Theoretical Analyses
1.2.5 From RSSI to CSI
2 Theoretical Model for RSS Localization
2.1 Radio Propagation Model
2.1.1 Theoretical Model
2.1.2 Challenges and Motivations
2.2 Accuracy and Reliability Analysis
2.2.1 System Model
2.2.2 Localization in One-Dimensional Space
2.2.2.1 One-Time Measurement for Single AP
2.2.2.2 Multiple Measurements for Multiple APs
2.2.2.3 Discussions
2.2.3 Localization in Two-Dimensional Space
2.2.3.1 Multiple Measurements for Multiple APs
2.2.3.2 Discussions
2.2.4 Best Strategy for Location Determination
2.2.4.1 Fundamentals of Location Determination
2.2.4.2 Best Strategy
2.2.5 Impact on Localization in One-Dimensional Space
2.2.5.1 Imperfect Information
2.2.5.2 One-Time Measurement for Single AP
2.2.5.3 Multiple Measurements for Multiple APs
2.2.6 Impact on Localization in Two-Dimensional Space
2.2.7 Experimental Results
2.2.7.1 Verification of Main Assumptions
2.2.7.2 Localization Performance
2.3 Scalability with the Collocation of Measurement Points
2.3.1 Case 1: Regular Collocation
2.3.1.1 Background and Modeling
2.3.1.2 Theoretical Analysis
2.3.1.3 Comparisons of Collocation Patterns
2.3.2 Random Collocation
2.3.2.1 Background and Modeling
2.3.2.2 Theoretical Analysis
2.3.3 Asymmetrical Distribution
2.3.3.1 Modeling
2.3.3.2 Theoretical Analysis
2.3.4 Simulation
2.3.4.1 Regular Collocation
2.3.4.2 Random Collocation
2.3.4.3 Asymmetrical Distribution
2.4 Scalability with the Number of Users
2.4.1 Problem Formulation
2.4.1.1 Interference Region
2.4.1.2 Localization Reliability Model
2.4.1.3 Localization Performance Deterioration by Human Body Blockage
2.4.1.4 Strategy of Deriving the Scalability
2.4.2 Localization Reliability Without Blockage
2.4.3 Localization Reliability Deterioration by Blockage
2.4.3.1 Localization Reliability with Blockage
2.4.3.2 Finding Bounds of R
2.4.3.3 Upper Bound of R
2.4.3.4 Lower Bound of R
2.4.3.5 Magnitude of R w.r.t. m
2.4.4 Number of Impacted Access Points w.r.t. Number ofUsers
2.4.4.1 Shape of Influence Region
2.4.4.2 Bounding the Number of Impacted APs
2.4.4.3 Determine the Environment Dependent Parameters
2.4.5 Main Results
2.4.6 Evaluations of Main Results
2.4.6.1 Influence of Human Body Blockage
2.4.6.2 Numerical Results
2.4.6.3 Experimental Results
2.4.6.4 Important Observations and Analysis
2.5 Theoretical Guidance on Fingerprints Reporting Strategy
2.5.1 System Model
2.5.2 Analysis of Optimal Strategy for Fingerprints Reporting
2.5.2.1 Supermodularity of the Objective Function
2.5.2.2 Algorithm for AP Selection
2.5.2.3 Algorithm Performance Analysis
2.5.3 Applications of the Best Strategy
2.5.3.1 Location Estimation Leveraging Best Fingerprints Reporting Strategy
2.5.3.2 Strategy for AP Deployment
2.5.4 Performance Evaluation
2.6 Theoretical Guidance on Optimizing Localization Accuracy
2.6.1 Theoretical Model of Location Estimation
2.6.2 Analysis of 2D Temporal Correlation for 1D Localization
2.6.2.1 Finding Region E
2.6.2.2 Analysis on Region E
2.6.2.3 Influence of Temporal Correlation on Accuracy of Localization
2.6.3 Asymptotic Equivalent Region of E in High-Dimensional Scenarios
2.6.3.1 Approximate Matrix
2.6.3.2 Asymptotical Equivalence Analysis
2.6.3.3 Boundaries of Region E
2.6.4 Location Estimation Facilitated by Temporal Correlation
2.6.4.1 Feasibility of Utilizing Temporal Correlation
2.6.5 Localization Estimation Algorithm
2.6.5.1 Choice of Design Parameters
2.6.6 Experimental Results
3 Theoretical Model for CSI Localization
3.1 System Model
3.1.1 Signal Model
3.1.2 Problem Formulation
3.1.3 Challenges
3.2 CRB Analysis in Frequency Domain
3.2.1 Parameter Analysis with Single Antenna
3.2.2 Parameter Estimation with Antenna Array
3.2.3 CRB for Location Estimation
3.3 Asynchronization Analysis
3.3.1 Parameter Estimation Over Single Antenna
3.3.2 Parameter Estimation Over Multiple Antennas
3.4 Insight into CSI Approach
3.5 Experimental Results
3.5.1 Asynchronization Error Bound
3.5.2 Antenna Array Design
3.5.3 Practical Localization Performance
4 RSS Localization for Large-Scale Deployment
4.1 Online Pricing Mechanism for Crowdsensing Localization Data
4.1.1 System Model
4.1.2 Quality Evaluation of Fingerprints
4.1.2.1 Probability Model of Data Error
4.1.2.2 Data Error Analysis
4.1.3 Quality-Aware Online Pricing Mechanism Design
4.1.3.1 Loss and Regret Function
4.1.3.2 Quality-Aware Online Pricing Scheme
4.1.4 Data Pricing with Budget Constraints
4.1.4.1 Regret Minimization with Fixed Budget
4.1.4.2 Budget Minimization for Certain Quality Level
4.1.5 Experimental Results
4.2 Incentive Mechanism for Mobile Crowd Sensing
4.2.1 System Model and Design Challenges
4.2.1.1 System Model
4.2.1.2 Design Challenges
4.2.2 Quality-Driven Auction
4.2.2.1 Overview
4.2.2.2 Particular Value of the Sub-contract
4.2.3 Algorithm of QDA
4.2.4 Proving Properties of QDA
4.2.5 Applying QDA to the Indoor Localization System
4.2.6 Performance Evaluation
4.2.6.1 Truthfulness and Individual Rationality
4.2.6.2 Social Welfare
4.2.7 Quality Discrimination
4.2.8 Computational Cost
4.3 Prediction for Fingerprints Data (Quadrotors)
4.3.1 Working Process of HiQuadLoc
4.3.1.1 Offline Data Training Phase
4.3.1.2 Online Localization Phase
4.3.1.3 Turning Detection
4.3.1.4 Structure of Localization Algorithm
4.3.2 Preliminary Localization Algorithm
4.3.2.1 Theoretical Basis
4.3.2.2 4-D RSS Interpolation Scheme in Offline Phase
4.3.2.3 Preliminary Position Estimation in Online Phase
4.3.3 Path Correction Scheme
4.3.3.1 Path Estimation
4.3.3.2 Parameter Readjustment During Turning
4.3.3.3 Path Fitting
4.3.3.4 Location Prediction
4.3.4 Experiment Results
4.3.4.1 Evaluation of 4-D RSS Interpolation Algorithm
4.3.4.2 Evaluation of Localization Schemes
4.3.4.3 Evaluation of Parameter Readjustment During Turning
4.3.4.4 Evaluation of HiQuadLoc for Different Flight Speeds
4.3.5 Comparison with Channel State Information (CSI) Based Scheme
4.4 Prediction for Fingerprints Data (Cellular Network)
4.4.1 Problem Formulation
4.4.1.1 Fingerprints Prediction: A Subspace Identification Perspective
4.4.1.2 Problem Formulation
4.4.2 Streamlined Stiefel Manifold Optimization
4.4.2.1 Algorithm Design
4.4.2.2 Convergence Analysis
4.4.2.3 Discussions
4.4.3 Fingerprints Prediction with Sliding Window
4.4.3.1 Sliding Window Mechanism Design
4.4.4 Remained Information Analysis
4.4.4.1 Sliding Window Algorithm
4.4.5 Experimental Results
4.4.5.1 Experiments on Small Data Set
4.4.5.2 Experiments on Large Data Set
4.5 Floor Plan Generation for Localization
4.5.1 Motivation
4.5.1.1 Analysis of Wi-Fi Landmarking
4.5.1.2 BLE Landmarking
4.5.2 System Overview
4.5.3 Data Collection
4.5.3.1 Trace Data Format
4.5.3.2 Posture Recognition in Dead Reckoning
4.5.4 Trace Labeling
4.5.4.1 Labeling Traces with BLE Beacons
4.5.4.2 Labeling Trace Segments in Rooms
4.5.4.3 Trace Merging
4.5.5 Trace Revising
4.5.6 Map Pixel Classification
4.5.7 Map Construction and Localization
4.5.8 Performance Evaluation
4.5.8.1 Experimental Setups
4.5.8.2 Accuracy of Posture Recognition
4.5.8.3 Performance of Map Construction
4.5.8.4 System Comparison
4.5.8.5 Localization Performance
4.5.9 Discussions
5 CSI Localization for Large-Scale Deployment
5.1 Extended MUSIC Algorithm
5.1.1 Preliminaries
5.1.1.1 OFDMA Backscatter System
5.1.1.2 Challenges
5.1.2 System Overview
5.1.3 OFDM Burst Processing
5.1.3.1 3D-CSI Analysis
5.1.3.2 OFDM Burst Processing
5.1.4 Phase Offset Elimination
5.1.4.1 Continuous Dynamic Phase Offset
5.1.4.2 Down Conversion Phase Offset
5.1.5 Extended MUSIC Scheme
5.1.5.1 Derive AoA of Tags with Limited Information
5.1.5.2 Symbol-Domain Extension
5.1.5.3 Multi-Domain Extension
5.1.6 Implementation and Concurrent Localization
5.1.7 Experiments
5.1.7.1 Necessity of System Designs
5.1.7.2 Localization Performance
5.1.8 Performance Comparison
5.2 Autonomous Wi-Fi Device Map
5.2.1 Motivation and Challenge
5.2.1.1 Motivation
5.2.1.2 Challenge
5.2.2 System Overview
5.2.3 Self-Calibrating System Designs
5.2.3.1 Basic Idea
5.2.3.2 Scenario A: Phase Distortion Spectrum Analysis
5.2.3.3 Scenario B: Triangulation Analysis
5.2.3.4 Layout Construction and Localization
5.2.4 Non-Linear Antenna Array Designs
5.2.4.1 Limitations of Linear Antenna Array
5.2.4.2 Model the Antenna Sub-Array
5.2.4.3 Model the Antenna Array
5.2.5 Performance Evaluation
5.2.5.1 Implementation
5.2.5.2 Performance of Different Antenna Layouts
5.2.5.3 Construction Performance
5.2.5.4 Localization Performance
5.2.5.5 NLoS Scenarios
5.2.5.6 System Comparison
5.3 Calibration-Free CSI Fingerprints
5.3.1 Analysis of Fingerprinting Localization
5.3.1.1 The Size of Cells
5.3.1.2 The Type of Fingerprints
5.3.2 System Overview
5.3.3 Theoretical Fingerprints Generation
5.3.3.1 Basic Idea and Challenges
5.3.3.2 Mapping AoA into Phase Difference
5.3.3.3 Generate Super-Resolution Fingerprints
5.3.4 Fingerprinting Localization
5.3.4.1 Single-Spot Localization
5.3.4.2 Euclidean Distance Multiplication
5.3.4.3 LSTM Network
5.3.5 Automatic Fingerprints Update
5.3.6 Performance Evaluation
5.3.6.1 Theoretical Fingerprints Database
5.3.6.2 Fingerprinting Localization
5.3.6.3 Automatic Fingerprints Update
5.3.6.4 System Comparison
6 Conclusions
6.1 Research Summary
6.2 Future Work
References
Index