Intercambio de datos con RISC-V en el ámbito IoT
Loading...
Official URL
Full text at PDC
Publication date
2025
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citation
Abstract
El crecimiento exponencial del Internet de las Cosas está expandiendo sus fronteras más allá del ámbito terrestre, dando lugar a innovadoras aplicaciones en el sector espacial. En este campo, las constelaciones de nanosatélites CubeSat de bajo coste para la observación de la Tierra se enfrentan a severas restricciones de ancho de banda y energía. Estas limitaciones convierten la transmisión indiscriminada de datos, como imágenes con alta cobertura de nubes, en un proceso ineficiente que malgasta recursos críticos para la misión. Para demostrar y validar la capacidad de la arquitectura RISC-V en el intercambio y procesamiento de datos en un entorno IoT exigente, este trabajo implementa un caso de uso de teledetección híbrida. Para ello, se propone una arquitectura basada en un System-on-Chip con un núcleo de procesador RISC-V, implementada sobre una plataforma FPGA (una Nexys A7). Sobre este sistema, se integra un módulo de comunicación Ethernet para simular el enlace satelital y se desarrolla el firmware en C necesario para ejecutar un algoritmo de detección de nubes basado en K-means. Las pruebas del sistema validan la prueba de concepto, demostrando la capacidad del procesador para analizar imágenes de 100x100 píxeles, descartar de forma autónoma aquellas con un porcentaje de opacidad superior al umbral del 20 %, y transmitir exitosamente las imágenes útiles a una estación terrestre simulada. Este trabajo concluye que la arquitectura RISC-V es una plataforma robusta y eficiente para el intercambio y procesamiento de datos en aplicaciones IoT complejas, como ha quedado demostrado con la exitosa implementación de un nodo de teledetección satelital. La capacidad del sistema para realizar cómputo en el borde y filtrar datos de forma autónoma valida su flexibilidad y rendimiento, consolidándola como una alternativa estratégica para el desarrollo de nodos IoT inteligentes que requieran optimizar su flujo de información.
The exponential growth of the Internet of Things (IoT) is expanding its frontiers beyond the terrestrial sphere, leading to innovative applications in the space sector. In this field, low-cost CubeSat nanosatellite constellations for Earth observation face severe bandwidth and power constraints. These limitations turn the indiscriminate transmission of data, such as images with high cloud coverage, into an inefficient process that wastes mission-critical resources. To demonstrate and validate the capability of the RISC-V architecture for data exchange and processing in a demanding IoT environment, this work implements a hybrid remote sensing use case. To this end, an architecture based on a System-onChip (SoC) with a RISC-V processor core is proposed, implemented on an FPGA platform (a Nexys A7). On this system, an Ethernet communication module is integrated to simulate the satellite link, and the necessary C firmware is developed to execute a K-means-based cloud detection algorithm. System tests validate the proof-of-concept, demonstrating the processor’s ability to analyze 100x100 pixel images, autonomously discard those with an opacity percentage above a 20% threshold, and successfully transmit the useful images to a simulated ground station. This work concludes that the RISC-V architecture is a robust and efficient platform for data exchange and processing in complex IoT applications, as has been demonstrated by the successful implementation of a satellite remote sensing node. The system’s ability to perform edge computing and autonomously filter data validates its flexibility and performance, establishing it as a strategic alternative for developing intelligent IoT nodes that need to optimize their information flow.
The exponential growth of the Internet of Things (IoT) is expanding its frontiers beyond the terrestrial sphere, leading to innovative applications in the space sector. In this field, low-cost CubeSat nanosatellite constellations for Earth observation face severe bandwidth and power constraints. These limitations turn the indiscriminate transmission of data, such as images with high cloud coverage, into an inefficient process that wastes mission-critical resources. To demonstrate and validate the capability of the RISC-V architecture for data exchange and processing in a demanding IoT environment, this work implements a hybrid remote sensing use case. To this end, an architecture based on a System-onChip (SoC) with a RISC-V processor core is proposed, implemented on an FPGA platform (a Nexys A7). On this system, an Ethernet communication module is integrated to simulate the satellite link, and the necessary C firmware is developed to execute a K-means-based cloud detection algorithm. System tests validate the proof-of-concept, demonstrating the processor’s ability to analyze 100x100 pixel images, autonomously discard those with an opacity percentage above a 20% threshold, and successfully transmit the useful images to a simulated ground station. This work concludes that the RISC-V architecture is a robust and efficient platform for data exchange and processing in complex IoT applications, as has been demonstrated by the successful implementation of a satellite remote sensing node. The system’s ability to perform edge computing and autonomously filter data validates its flexibility and performance, establishing it as a strategic alternative for developing intelligent IoT nodes that need to optimize their information flow.
Description
Trabajo de Fin de Máster en Internet de las Cosas, Facultad de Informática UCM, Departamento de Arquitectura de Computadores y Automática, Curso 2024/2025.