Experimental Demonstration of 3D Visible Light Positioning Using Received Signal Strength With Low-Complexity Trilateration Assisted by Deep Learning Technique

Pengfei Du,Chen Chen,Helin Yang,Wen-De Zhong,Yanbing Yang,Sheng Zhang,Ran Zhang,Arokiaswami Alphones

doi:10.1109/access.2019.2928014

Abstract

In this paper, a 3D indoor visible light positioning (VLP) system with fast computation time using received signal strength (RSS) is proposed and experimentally demonstrated. Assisted by the deep learning techniques, the complexity of the trilateration problem is greatly reduced, and the trilateration problem can be formulated as a linear mapping leading to faster position estimation than the conventional estimation. Moreover, a new method of off-line preparation is adopted to minimize the workload of the VLP system deployment for more practical usage. The proposition is implemented on an atto-cellular VLP unit, through which the real-time performance and positioning accuracy are demonstrated and validated in a 3D positioning experiment performed in a space of 1.2 × 1.2 × 2 m 3 . The experimental results show that a positioning accuracy of 11.93 cm in confidence of 90% is achieved with 50 times faster the computation time compared to the conventional scheme.

Highlights

Within recent years, indoor visible light positioning (VLP) services based on visible light communication (VLC) systems become more and more attractive to researchers as the light-emitting diodes (LED) lighting infrastructure is being deployed worldwide [1], [2]
In this paper, a practical 3D VLP system using received signal strength (RSS) with lowcomplexity trilateration assisted by deep learning is proposed to address: 1) the complicated solution to the trilateration problem; 2) the shortcoming of the previous existing scheme relying on tedious offline preparation
The proposed VLP system is elaborated by separately introducing the channel modeling, the Lambertian order measurement in offline preparation and the online position estimation simplified by deep learning technique along with theoretical derivations

Summary

INTRODUCTION

Indoor visible light positioning (VLP) services based on visible light communication (VLC) systems become more and more attractive to researchers as the LED lighting infrastructure is being deployed worldwide [1], [2]. Some researchers adopt machine-learning techniques to realize fingerprinting based 3D VLP system using RSS so as to completely circumvent the trilateration calculation Such a system has tedious offline workload and terrible user friendliness as stated above [19], [28]. In order to address the challenges above, this paper proposes and experimentally demonstrates a practical 3D VLP system, which retains the distance estimation using RSS and combines machine learning to simplify the trilateration solution so that the computation speed can be substantially increased with minimum offline preparation work. PRINCIPLE This section introduces the principle of the proposed VLP system, including the overall system model, RSS-based distance estimation and simplified trilateration solution assisted by deep learning techniques. The position is given by the position estimator based on the signal attenuation model and the trilateration method, whereas the deep learning technique is adopted for performance enhancement

DISTANCE ESTIMATION BASED ON RSS

Findings

CONCLUSION