Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Machine Learning-Based Direct Source Localization for Passive Movement-Driven Virtual Large Array

Shang-Ling Shih, Chao-Kai Wen, Chau Yuen, Shi Jin

TL;DR

This paper introduces a novel smartphone-enabled localization technology for ambient Internet of Things (IoT) devices, leveraging the widespread use of smartphones, that enables direct localization using only angle-of-arrival (AoA) information.

Abstract

This paper introduces a novel smartphone-enabled localization technology for ambient Internet of Things (IoT) devices, leveraging the widespread use of smartphones. By utilizing the passive movement of a smartphone, we create a virtual large array that enables direct localization using only angle-of-arrival (AoA) information. Unlike traditional two-step localization methods, direct localization is unaffected by AoA estimation errors in the initial step, which are often caused by multipath channels and noise. However, direct localization methods typically require prior environmental knowledge to define the search space, with calculation time increasing as the search space expands. To address limitations in current direct localization methods, we propose a machine learning (ML)-based direct localization technique. This technique combines ML with an adaptive matching pursuit procedure, dynamically generating search spaces for precise source localization. The adaptive matching pursuit minimizes location errors despite potential accuracy fluctuations in ML across various training and testing environments. Additionally, by estimating the reflection source's location, we reduce the effects of multipath channels, enhancing localization accuracy. Extensive three-dimensional ray-tracing simulations demonstrate that our proposed method outperforms current state-of-the-art direct localization techniques in computational efficiency and operates independently of prior environmental knowledge.

Machine Learning-Based Direct Source Localization for Passive Movement-Driven Virtual Large Array

TL;DR

This paper introduces a novel smartphone-enabled localization technology for ambient Internet of Things (IoT) devices, leveraging the widespread use of smartphones, that enables direct localization using only angle-of-arrival (AoA) information.

Abstract

This paper introduces a novel smartphone-enabled localization technology for ambient Internet of Things (IoT) devices, leveraging the widespread use of smartphones. By utilizing the passive movement of a smartphone, we create a virtual large array that enables direct localization using only angle-of-arrival (AoA) information. Unlike traditional two-step localization methods, direct localization is unaffected by AoA estimation errors in the initial step, which are often caused by multipath channels and noise. However, direct localization methods typically require prior environmental knowledge to define the search space, with calculation time increasing as the search space expands. To address limitations in current direct localization methods, we propose a machine learning (ML)-based direct localization technique. This technique combines ML with an adaptive matching pursuit procedure, dynamically generating search spaces for precise source localization. The adaptive matching pursuit minimizes location errors despite potential accuracy fluctuations in ML across various training and testing environments. Additionally, by estimating the reflection source's location, we reduce the effects of multipath channels, enhancing localization accuracy. Extensive three-dimensional ray-tracing simulations demonstrate that our proposed method outperforms current state-of-the-art direct localization techniques in computational efficiency and operates independently of prior environmental knowledge.

Paper Structure

This paper contains 23 sections, 1 theorem, 55 equations, 9 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model
  3. Coordinate System
  4. Array Response Before Moving and Rotation
  5. Array Response After Moving and Rotation
  6. Signal Model for the VLA
  7. Performance Bounds
  8. FIM of VLA
  9. FIM of RLA
  10. Comparison of RLA and VLA
  11. Localization Algorithm
  12. Direct Localization via NOMP
  13. Search Space Initialization for PS
  14. Search Space Initialization for VSs
  15. Summary of ML-NOMP Algorithm
  16. ...and 8 more sections

Key Result

Proposition 1

The FIM difference between the RLA and the VLA is given by where and $\tilde{\qp}_i = \qp - \qdelta_i$. The trace of the $\Delta {\rm FIM}_I$ is given by where is the angle between vectors $\tilde{\qp}_i$ and $\tilde{\qp}_j$.

Figures (9)

  • Figure 1: Locations of the PS, VS, and the moving UE.
  • Figure 2: The location candidates of $\qp$ determined by the angle information obtained at $\qo$ (purple line), the angle information obtained at $\qdelta$ (orange line), and the range difference information (blue line) for different values of $\rho$.
  • Figure 3: The SPEB of RLA and VLA with different values of $\|\qdelta_I\|$.
  • Figure 4: Search space determined for different sources by various methods: (a) applying the traditional method for PS, (b) applying the ML method for PS, (c) utilizing estimated distance for VS.
  • Figure 5: Input data for a single trajectory.
  • ...and 4 more figures

Theorems & Definitions (1)

  • Proposition 1