Contribution to Optical Camera Communication using multispectral cameras and LED’s thermal effect

Abstract

Optical wireless communication (OWC) is progressively emerging as a pivotal technology in upcoming communication networks and is believed to be poised for a revolution within both industry and research sectors. Its capacity to offer huge bandwidth, rapid deployment, and low power consumption has attracted significant attention. Hence, these characteristics position OWC as a compelling candidate to meet the requirements of future communication systems. Among the principal technologies encompassing OWC, optical camera communication (OCC) has experienced a pronounced increase in recent years owing to the proliferation of camera-equipped devices. Consequently, the widespread adoption of these devices, along with the advances in image sensors, has boosted the interest of OCC. This technology overcomes the limitation of visible light communication (VLC) and light fidelity (LiFi) against path loss and eases the complexity of the implementation with respect to free-space optical (FSO) links. However, one of the major challenges of OCC lies in increasing the data rate, which has led to the development of new techniques and approaches to improve it. In this context, the use of non-conventional cameras is leveraged in order to improve the OCC system performance. Given that the data rate is partially defined by half of the camera’s frame rate, high-speed cameras (more than 120 fps) are the predominant choice for increasing the data rate. Despite that constraint, certain applications prioritize factors other than data rate, making them well-suited for the incorporation of different types of camera to enhance overall system performance. To the best of the author’s knowledge, the use of high-spectral-resolution cameras as receivers in OCC systems remains relatively unexplored. Multispectral (MS) and hyperspectral (HS) cameras provide more bands than conventional red, green, and blue (RGB) types by using narrow band-pass filters. The number of bands, as well as their spectral width, determines the camera’s spectral resolution. Those characteristics make these kinds of cameras a valuable tool in several applications, such as remote sensing, agriculture, and medicine. Specifically, MS and HS cameras are set to become a vital factor in precision farming due to the increasing demands in food production, making it necessary to optimize farming practices sustainably. This considerable interest and the particular features of these devices open up a variety of opportunities for communication. Notably, this operating principle of this type of cameras permits capturing the spectral response curve, also known as spectral signature, of different elements of an image. Thus, using light-emitting diodes (LEDs) with different peak wavelengths transmitting certain data, the MS camera captures the signals and could separate them taking advantage of their distinct spectral signature, to decode the information ultimately. On the other hand, the employment of LEDs has considerably increased in the last decade. Developments in this type of device have caused improvements in their efficiency, allowing high luminous intensity with low power consumption. As a result, it has raised the LED demands in several sectors, such as industrial, commercial, and residential. This increase has also heightened the OWC attention because they are customarily used as the transmitter light source. The key feature that makes these devices helpful for communication is their rapid switching capability that allows modulating light. Moreover, LEDs are sensitive to temperature variations that affect their performance and spectral features. In general, luminous efficiency decreases, and the peak wavelength increases as temperature grows. Even though the temperature dependence of LEDs is well-known, few researchers have addressed the impact of thermal effects in OWC. Nonetheless, this effect is analyzed in this thesis. On the one hand, it was suggested to increase the communication channels that can be attained using a single device by taking advantage of the increase in the LED’s wavelength (red-shift) along with using a multispectral camera to capture the spectral variations. Thus, this reaction, typically considered harmful to the system performance, is turned into a benefit. On the other hand, the consequences of not taking LED spectral variations in approaches that require channel compensation to diminish inter-channel interference (ICI) are examined. In this thesis, a set of objectives has been formulated to study the use of multispectral cameras and LED thermal effects in OCC. First, it is intended to implement an OCC link based on the LED’s temperature effects and a multispectral camera. Multispectral cameras possess a distinct advantage in their high-spectral resolution, which makes them suitable candidates for capturing and differentiating the spectral variations induced by thermal changes in LEDs. By implementing an OCC link that capitalizes on LED temperature effects and utilizes MS cameras, the objective is to demonstrate the feasibility of establishing multiple communication channels from a single light source. Second, it is intended to implement nonlinear techniques for data detection in multispectral camera communication (MCC) systems with the purpose of enhancing the system performance. In this context, the primary goal lies in a comparative analysis between the performance of the OCC system employing these nonlinear methodologies and that of conventional linear methods. Finally, the objective is to analyze thermal effects on LEDs in OCC systems’ compensation stage. This objective centers on a comprehensive analysis of the thermal effects on LEDs in OCC systems using conventional RGB cameras. The goal is to quantify the impact of temperature-induced spectral variations on system performance and establish the necessity of compensating for these effects. In order to achieve these objectives, the basic methodology followed throughout this thesis involves three main steps. First, analyze the LED behavior when its p-n junction temperature changes by developing a mechanism for controlling the temperature. Second, collect behavior characteristics from the light sources using various instrumentation equipment to characterize them and an MS camera to establish a communication link. Lastly, process data by applying several techniques to analyze them. In this thesis, both simulations and experimental studies have been conducted. On the one hand, the thermal impact of LEDs on channel compensation in the context of OCC has been addressed. Thus, an OCC link employing an RGB LED was assessed when obsolete channel state information (CSI) in terms of temperature is used for the channel compensation, i.e., when the actual p-n junction temperature of the transmitter is different from the LED temperature at optimal working conditions from which the CSI was estimated. Therefore, several temperatures were induced to the light source under test based on the Joule effect by increasing the driving current of the device. Both temperature and LED emitted signals were characterized using a thermal camera and an optical spectrometer, respectively. Then, the responses of Bayer-based and Foveon image sensors were simulated to obtain the associated CSI and perform the zero-forcing (ZF) compensation of two channel matrices at different temperatures. Finally, the system performance was evaluated, estimating signal-to-interference-plusnoise ratio (SINR) and bit error rate (BER), and demonstrating performance degradation due to LED spectral variations. Moreover, an innovative configuration using an MS camera as a receiver in an OCC system and the thermally induced spectral changes of LEDs was carried out. The effects of temperature on LEDs, which are usually considered a detrimental factor, were used in this study to increase the number of communication channels using the same device. To achieve this result, an on-off keying (OOK) transmission was simulated to analyze the system performance. It consisted of generating a transmitted signal (a bit stream after being affected by the channel matrix) to which noise was added. Then, a ZF equalizer was used to estimate the transmitted bit stream, applying the Moore-Penrose pseudoinverse of the channel matrix. Once the transmitted bit stream was estimated, it was compared to the sent bit stream to calculate the BER. Finally, considering the MS camera and the effect of temperature on LEDs, a cluster-based data detection approach was performed for improving the performance of an OCC system. The balanced iterative reducing and clustering hierarchies (BIRCH) algorithm was used to generate a clustering model with the purpose of recovering the LED’s signals. This cluster analysis results in a BER enhancement with respect to linear methods, namely, ZF and minimum mean square error (MMSE). In addition, this experimental study proposed a novel approach exploring the possibility of adopting a spectral signature multiplexing based on temperature. The success of the approach that integrates high-spectral-resolution cameras with the consideration of temperature effects on LEDs hinges on several key factors. Primarily, the spectral variations that LEDs present due to temperature fluctuations are contingent upon the materials composing the LEDs and the characteristics of the LED substrate. Typically, as temperature increases, the peak wavelength of LEDs tends to shift towards longer wavelengths, albeit at the cost of reduced efficiency. Striking a balance between accommodating this red-shift and mitigating the accompanying reduction in efficiency is of critical importance. Concurrently, the effectiveness of this approach is intrinsically linked to the spectral resolution of the camera employed. A camera’s spectral resolution dictates its sensitivity to spectral variations, and as such, cameras with a greater number of spectral bands and narrower bandpass widths are better equipped to capture and discern such variability. Nevertheless, the adoption of this technology may encounter certain limitations. Mainly, the high acquisition cost of high-spectral-resolution cameras, making integration into commercial off-the-shelf (COTS) devices impractical, and limited data rates due to low frame rates in most of these cameras. Despite these challenges, some models offer rolling shutter (RS) mode for increased data rates, but the non-communication-oriented design of cameras may lead to issues such as data loss or data stream merging. Additionally, addressing temperature-induced spectral changes in LEDs is crucial for maintaining reliable signal demultiplexing. Considering the constraints and potential attributes of high-spectralresolution cameras for communication purposes, several compelling applications can be envisioned. To mitigate the expense associated with acquiring such cameras, their primary utility should focus on enhancing existing systems equipped with cameras for communication tasks. For example, in industrial automation and quality control, these cameras can serve a dual role by detecting contaminants in food processing lines and monitoring the performance of conveyor systems, transmitting this information via OCC links. This approach facilitates real-time monitoring and maintenance in production facilities. Similarly, in precision agriculture, multispectral cameras can provide early disease detection, parasite presence, and critical data on soil nutrient levels and moisture content. Placing these cameras on fixed stands within agricultural fields offers dual benefits of crop health monitoring and data reception from field-installed sensors, facilitating informed decision-making. These applications highlight the multifaceted utility of high-spectral-resolution cameras, enhancing efficiency, monitoring, and decision-making across various domains.

La comunicación óptica inalámbrica (OWC) se perfila de manera progresiva como una tecnología crucial en las futuras redes de comunicación y se cree que está preparada para una revolución en la industria y el ámbito de la investigación. Su capacidad para ofrecer un ancho de banda amplio, una implementación rápida y un bajo consumo de energía ha atraído una atención significativa. Por lo tanto, estas características sitúan a la OWC como un candidato válido para satisfacer los requisitos de los futuros sistemas de comunicación. Dentro de las principales tecnologías que abarcan la OWC, la comunicación óptica a través de cámaras (OCC) ha experimentado un notable crecimiento en los últimos años debido a la proliferación de dispositivos equipados con cámaras. Como resultado, la adopción generalizada de estos dispositivos, junto con los avances en los sensores de imagen, ha impulsado el interés de la OCC. Esta tecnología supera la limitación de la comunicación por luz visible (VLC) y LiFi frente a la pérdida de señal y simplifica la complejidad de la implementación en comparación con las comunicaciones ópticas por el espacio libre (FSO). Sin embargo, uno de los principales desafíos de la OCC radica en aumentar la velocidad de transmisión de datos, lo que ha llevado al desarrollo de nuevas técnicas y enfoques para mejorarla. En este contexto, se aprovecha el uso de cámaras no convencionales para mejorar el rendimiento del sistema de OCC. Dado que la velocidad de transmisión de datos está parcialmente definida por la mitad de la tasa de fotogramas de la cámara, las cámaras de alta velocidad (más de 120 fotogramas por segundo) son la elección predominante para incrementar la tasa de datos. A pesar de esa limitación, ciertas aplicaciones priorizan factores diferentes a la velocidad de transmisión de datos, lo que las hace adecuadas para la incorporación de diferentes tipos de cámaras con el fin de mejorar el rendimiento general del sistema. Dada la escasa literatura en este ámbito, el uso de cámaras de alta resolución espectral como receptores en sistemas de OCC aún se encuentra poco explorado. Las cámaras multiespectrales (MS) e hiperespectrales (HS) proporcionan más bandas que las roja, verde y azul (RGB) convencionales mediante el uso de filtros de paso de banda estrechos. El número de bandas, así como su anchura espectral, determina la resolución espectral de la cámara. Estas características hacen que este tipo de cámaras sea una herramienta valiosa en diversas aplicaciones, como la teledetección, la agricultura y la medicina. Específicamente, las cámaras MS y HS se están convirtiendo en un factor vital en la agricultura de precisión debido a las crecientes demandas en la producción de alimentos, lo que hace necesario optimizar las prácticas agrícolas de manera sostenible. Este considerable interés y las características particulares de estos dispositivos abren una variedad de oportunidades para la comunicación. Es importante destacar que el principio de funcionamiento de este tipo de cámaras permite capturar la curva de respuesta espectral, también conocida como firma espectral, de diferentes elementos de una imagen. De esta manera, utilizando diodos emisores de luz (LEDs) con diferentes longitudes de onda máximas para transmitir ciertos datos, la cámara MS captura las señales y podría separarlas aprovechando su distintiva firma espectral para finalmente decodificar la información. Por otro lado, el uso de LEDs ha aumentado considerablemente en la última década. Los avances en este tipo de dispositivos han llevado a mejoras en su eficiencia, lo que permite una alta intensidad luminosa con un bajo consumo de energía. Como resultado, ha aumentado la demanda de LEDs en varios sectores, como el industrial, comercial y residencial. Este aumento también ha incrementado la atención de la OWC, ya que estos se utilizan comúnmente como fuente de luz transmisora. La característica clave que hace que estos dispositivos sean útiles para la comunicación es su capacidad de conmutación rápida que permite modular la luz. Además, los LEDs son sensibles a las variaciones de temperatura que afectan su rendimiento y características espectrales. En general, la eficiencia luminosa disminuye y la longitud de onda máxima aumenta a medida que aumenta la temperatura. A pesar de que la dependencia de la temperatura de los LEDs es un fenómeno bien conocido, pocos investigadores han abordado el impacto de los efectos térmicos en OWC. Sin embargo, este efecto se analiza en esta tesis. Por un lado, se sugirió aumentar los canales de comunicación que se pueden obtener utilizando un solo dispositivo aprovechando el incremento en la longitud de onda de los LEDs (desplazamiento hacia el rojo) junto con el uso de una cámara multiespectral para capturar las variaciones espectrales. Así, esta reacción, que generalmente se considera perjudicial para el rendimiento del sistema, se convierte en un beneficio. Por otro lado, se examinan las consecuencias de no tener en cuenta las variaciones espectrales de los LEDs en enfoques que requieren compensación de canal para reducir la interferencia entre canales (ICI). Las cámaras multiespectrales presentan características únicas con respecto a las cámaras convencionales. Así, se pretende explotar esas características específicas para alcanzar resultados que otros tipos de cámaras no podrían. Además, las variaciones espectrales en los LEDs debidas a cambios en la temperatura de la unión p-n, que normalmente es un efecto perjudicial para la comunicación, se considera una mejora para alcanzar nuevos canales de comunicación.