Smartphone-Based Indoor Map Construction: Principles and Applications (SpringerBriefs in Computer Science)

This book focuses on ubiquitous indoor localization services, specifically addressing the issue of floor plans. It combines computer vision algorithms and mobile techniques to reconstruct complete and accurate floor plans to provide better location-based services for both humans and vehicles via commodity smartphones in indoor environments (e.g., a multi-layer shopping mall with underground parking structures). After a comprehensive review of scene reconstruction methods, it offers accurate geometric information for each landmark from images and acoustics, and derives the spatial relationships of the landmarks and rough sketches of accessible areas with inertial and WiFi data to reduce computing overheads. It then presents the authors’ recent findings in detail, including the optimization and probabilistic formulations for more solid foundations and better robustness to combat errors, several new approaches to promote the current sporadic availability of indoor location-based services, and a holistic solution for floor plan reconstruction, indoor localization, tracking, and navigation. The novel approaches presented are designed for different types of indoor environments (e.g., shopping malls, office buildings and labs) and different users. A valuable resource for researchers and those in start-ups working in the field, it also provides supplementary material for students with mobile computing and networking backgrounds. Preface 6 Contents 7 1 Introduction of Indoor Map Construction 10 1.1 Introduction 10 Reference 11 2 Indoor Map Construction via Mobile Crowdsensing 12 2.1 Introduction 12 2.2 Design Overview 14 2.3 Landmark Modeling 15 2.3.1 The Landmark Model 15 2.3.2 Coordinates of Geometric Vertices 17 2.3.3 Connecting Points of Wall Segments 18 2.3.4 Example 18 2.4 Landmark Placement 19 2.4.1 Notations 19 2.4.2 Spatial Relation Acquisition 19 2.4.3 Problem Formulation 21 2.4.4 Optimization Algorithm 22 2.5 Map Augmentation 23 2.5.1 Wall Reconstruction 23 2.5.2 Hallway Reconstruction 26 2.5.3 Room Reconstruction 27 2.6 Connection Area Detection 28 2.6.1 Types of Connection Areas 28 2.6.2 Features 30 2.6.3 Unsupervised Classification 30 2.6.4 Refinement and Placement 31 2.6.5 Types of Connection Areas 31 2.7 Performance 33 2.8 Discussion 35 2.9 Related Work 36 2.10 Conclusion 37 References 38 3 Incremental Indoor Map Construction with a Single User 40 3.1 Introduction 40 3.2 Overview 42 3.3 Localization via a Single Image 42 3.4 Trajectory Calibration and Cleaning 45 3.4.1 Trajectory Calibration 46 3.4.2 Trajectory Cleaning 47 3.5 Map Fusion Framework 48 3.5.1 Dynamic Bayesian Network 48 3.5.2 Particle Filter Algorithm 49 3.6 Landmark Recognition 51 3.7 Compartment Estimation 52 3.8 Performance 54 3.9 Discussion 55 3.10 Related Work 56 3.11 Conclusion 57 References 57 4 Indoor Localization by Photo-Taking of the Environment 60 4.1 Introduction 60 4.2 Relative Position Measurement 63 4.3 Triangulation Method 66 4.3.1 User Operations and Location Computation 66 4.3.2 Criteria for Users to Choose Reference Objects 68 4.3.3 Robustness of the Localization Primitive 69 4.4 Site Survey for Reference Objects Coordinates 72 4.4.1 Location Estimation in Unmapped Environments 73 4.4.2 Experiments on Site Survey 74 4.5 Identifying Chosen Reference Objects 75 4.5.1 System Architecture and Workflow 76 4.6 Benchmark Selection of Reference Objects 78 4.6.1 Benchmark Selection Problem 78 4.6.2 NP-Completeness Proof 79 4.6.3 A Heuristic Algorithm 80 4.7 Improve Localization with Geographical Constraints 81 4.7.1 Experiment Results and Problems in Early Prototype 82 4.7.2 Geographical Constraints 83 4.7.3 System Localization Performance 84 4.8 Discussion 85 4.9 Related Work 86 4.10 Conclusion 87 References 88 5 Smartphone-Based Real-Time Vehicle Tracking in Indoor Parking Structures 89 5.1 Introduction 89 5.2 Design Overview 92 5.3 Trajectory Tracing 93 5.3.1 Conventional Approaches 93 5.3.2 Shadow Trajectory Tracing 94 5.3.3 Equivalence Proof 96 5.4 Real-Time Tracking 100 5.4.1 Intuition 100 5.4.2 Road Skeleton Model 101 5.4.3 Probabilistic Tracking Framework 102 5.4.4 Tracking Algorithms 103 5.5 Landmark Detection 106 5.5.1 Types of Landmarks 106 5.5.2 Feature and Classification Algorithm 108 5.5.3 Prediction and Rollback 109 5.6 Performance 109 5.7 Discussion 112 5.8 Related Work 113 5.9 Conclusions 115 References 115 Front Matter ....Pages i-ix Introduction of Indoor Map Construction (Ruipeng Gao, Fan Ye, Guojie Luo, Jason Cong)....Pages 1-2 Indoor Map Construction via Mobile Crowdsensing (Ruipeng Gao, Fan Ye, Guojie Luo, Jason Cong)....Pages 3-30 Incremental Indoor Map Construction with a Single User (Ruipeng Gao, Fan Ye, Guojie Luo, Jason Cong)....Pages 31-50 Indoor Localization by Photo-Taking of the Environment (Ruipeng Gao, Fan Ye, Guojie Luo, Jason Cong)....Pages 51-79 Smartphone-Based Real-Time Vehicle Tracking in Indoor Parking Structures (Ruipeng Gao, Fan Ye, Guojie Luo, Jason Cong)....Pages 81-109