BR-102024018238-A2 - Visible Light Location System for Indoor Environments

Abstract

This invention describes an indoor localization system that uses light signals generated by LED luminaires, applying visible light positioning integrated with an Artificial Neural Network (ANN) to determine position with high precision. Although technologies such as Bluetooth, Wi-Fi, LiDARs, and radars are alternatives for indoor localization, each operates on different principles and faces specific limitations. The technology developed in this invention stands out for using a training process based on an illuminance estimator, which simulates the data necessary for training the ANN, minimizing the need for experimental data collection. This method significantly increases the scalability and generalizability of the proposal, making it a robust and efficient alternative in a context where there is still no universally accepted technology considered the best for indoor localization.

Inventors

• GUILHERME MARCIO SOARES
• MATEUS RABELO FONSECA DO NASCIMENTO

Assignees

• UNIVERSIDADE FEDERAL DE JUIZ DE FORA - UFJF

Dates

Publication Date

20260317

Application Date

20240904

Claims (4)

1. 1. VISIBLE LIGHT LOCATION SYSTEM FOR INDOOR ENVIRONMENTS, characterized by comprising: a) Luminaires (1) composed of a transmitter circuit (2) and an array of LEDs (3). b) Receiver (4) to capture the modulated light signals and process them.
2. 2. VISIBLE LIGHT LOCATION SYSTEM FOR INDOOR ENVIRONMENTS, characterized by having the following features in its transmitter circuit: a) Microcontroller sends control signal (6) to modulate the state of transistor Q2 (7); b) Transistor Q1 (9) is used to control the peak current of the LED array (3), which can be defined by choosing resistor Re (11); c) The switched circuit modulates the luminous flux of the LED array (3) at a single frequency pre-established from the control signal (6); d) The circuit can be powered by a DC bus or by AC power supply in conjunction with a rectifier and a switched converter, such as a Buck converter, to perform the conversion into regulated DC voltage (5);
3. 3. VISIBLE LIGHT LOCATION SYSTEM FOR INDOOR ENVIRONMENTS, characterized by comprising the following steps in its receiver circuit: a) Received illuminance signal is converted into a current signal by means of a photodetector (12); b) Transimpedance amplifier (13) transforms the current signal into a voltage signal; c) Voltage signal is converted into a digital signal by means of an analog-to-digital converter (ADC - 14); d) Inside the microcontroller (15), the light signal feature extractor algorithm is executed and a pre-trained artificial neural network is executed, capable of determining the position of the receiver in the environment from the characteristics of the received light signal.
4. 4. VISIBLE LIGHT LOCATION SYSTEM FOR INDOOR ENVIRONMENTS, characterized by the fact that the artificial neural network training process is based on an illuminance estimator, which comprises the following steps: a) Calculation of direct and indirect illuminance at each point in the environment using luminous distribution files provided by the luminaire manufacturer. b) Creation of the database through illuminance simulations; c) Introduction of uncertainties in the sensor during training; d) Correlation of the characteristics of the luminous signals with the positions in the environment.

Description

TECHNICAL FIELD [01] The present invention relates to a location/positioning system that uses light signals generated by LED luminaires and is implemented in indoor environments, being contemplated in the visible light communication sector. The technology solves the problem of location in indoor environments, since currently there are several technologies that can be used for this purpose, such as Bluetooth, Wi-Fi signal, lidars and sonars, but there is no technology that is universally accepted as the best for the purpose of location, each having advantages and disadvantages compared to the others. STATE OF THE ART [02] The present invention differs substantially from previous work by introducing a technique that combines received signal strength (RSS) with fingerprinting, implemented through an Artificial Neural Network (ANN). Unlike traditional fingerprinting approaches, which are limited to offline and online phases to collect RSS data and determine the receiver's position in real time, the present invention goes further by using simulations based on IES files provided by luminaire manufacturers, eliminating the need for field data collection. This not only makes the system more generic and scalable, but also improves location accuracy, something that is not achieved by traditional techniques. [03] The works of Nadeem et al. (2015) and Luo et al. (2016) use RSS-based Fingerprinting, but without integration with an ANN, which limits the system's adaptability to different environments. Wei and Yao (2017) and Chen et al. (2018) also follow this approach, using Fingerprinting, but without exploring the potential of neural networks to improve the system's robustness. Zhao et al. (2017), in turn, propose a combination of Trilateration with Fingerprinting, offering a hybrid approach that still depends on data collection in the field, which may limit its applicability in practical situations. On the other hand, Oh and Kim (2022) and Chakraborty et al. (2022) explore techniques that combine RNA with Digital Printing, aligning more closely with the current proposal, but still without using simulations with IES files to eliminate the need for data collection in the field and without presenting circuits for the practical implementation of the proposal, thus limiting themselves to the description of the methodologies. [04] The articles by Lopes et al. (2019) and Do Nascimento et al. (2022) discuss the VLC localization technique in a basic way, focusing on theory and methodology, but without advancing to an experimental implementation. The article by Do Nascimento et al. (2024) offers a more detailed description of the technique, but remains in the theoretical field, without presenting practical results. [05] Other works, such as the article by Marson et al. (2019), focus on the development of transmitter circuits for VLC, aimed exclusively at communication, without exploring the application of these circuits in location systems. The article by Carneiro et al. (2019) describes the development of a receiver for VLC, which, although used as a basis for the receiver of the present invention, does not address its application in an integrated location system. The dissertation by Do Nascimento (2023) also delves into the theory and methodology of visible light localization, but does not present the circuits for implementing the proposed technique. [06] In contrast, the present invention addresses the complete physical system for implementing visible light localization technology. The invention incorporates an advanced illuminance estimator, capable of estimating the illuminance waveform over time through static and dynamic illuminance calculations, using classical photometry formulas. This innovation allows the creation of a robust database for training the ANN, with the introduction of controlled uncertainties in the sensor, increasing the system's robustness against practical variations. The references cited above are described below. [07] U. Nadeem, N.U. Hassan, M. A. Pasha, and C. Yuen. "Indoor positioning system designs using visible LED lights: performance comparison of TDM and FDM protocols." Electronics Letters 51(1):72-74, Jan 2015. DOI: 10.1049/el.2014.1668. [08] Zhijie Luo, WeiNan Zhang, and GuoFu Zhou. "Improved spring model based collaborative indoor visible light positioning." Optical Review 23(3):479- 486, Mar 2016. DOI: 10.1007/s10043-016-0204-z. [09] Hongtao Wei and Hao Yao. "Indoor visible light location algorithm based on virtual fingerprint database." In 2017 IEEE 2nd Advanced Information Technology, Electronic and Automation Control Conference (IAEAC). IEEE, Mar 2017. DOI: 10.1109/iaeac.2017.8054455. [010] Guo Chen, Shao Jian-Hua, Ke Wei, and Zhang Chun-Yan. "A visible indoor light positioning algorithm based on fingerprint." In 2018 4th Annual International Conference on Network and Information Systems for Computers (ICNISC). IEEE, Apr 2018. DOI: 10.1109/icnisc.2018.00022. [011] Chuhan Zhao, Hongming