| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 18.31 MB | Adobe PDF |
Authors
Abstract(s)
Ao longo dos últimos anos, a fiabilidade e a segurança dos sistemas embebidos utilizados em áreas críticas, como a saúde e o sector automóvel, têm suscitado um interesse crescente na comunidade científica e ganho maior consciencialização entre o público em geral. Esta tese aborda a modelação e a implementação de uma arquitetura software fiável e segura para um sistema embebido focado na aquisição e processamento de sinais fisiológicos, em particular o eletrocardiograma (ECG). O trabalho realizado visou o CardioWheel, um projeto em curso desenvolvido pela CardioID Technologies, destinado a aplicações nas áreas da saúde e do automóvel. As particularidades destas áreas quanto aos seus requisitos de segurança e proteção dos utilizadores servem de caso de estudo para mostrar as vantagens da arquitetura desenvolvida. Assim, no estudo realizado foi feito o levantamento dos requisitos do sistema que foram utilizados para projetar a máquina de estados da arquitetura em UML, a qual foi validada formalmente utilizando a ferramenta Uppaal e o modelo de autómatos finitos temporizados. Também foi feita uma análise de ameaças à arquitetura para validar os aspetos relacionados com a segurança. A arquitetura foi desenvolvida para microcontroladores ESP32 usando o ecossistema ESP-IDF e o FreeRTOS, para o que foram consideradas camadas independentes de hardware. A camada de comunicação é baseada no protocolo Bluetooth Low Energy (BLE) e permite a transmissão dos dados do nó final para um gateway e, posteriormente, para um servidor na nuvem. A operação de atualização de firmware usando o componente Over-The-Air (OTA) foi também implementada e validada quanto à sua segurança. A arquitetura foi, inicialmente, avaliada e validada usando um protótipo laboratorial. Posteriormente, foi utilizada para realizar uma pequena série de produção do CardioWheel em que se utilizaram as estratégias de validação propostas no contexto do projeto ESCEL KDT Valu3s. Também foi realizado um ensaio pré-médico no Hospital de Santa Marta usando o CardioWheel com a arquitetura proposta, que permitiu validar a sua fiabilidade e capacidades quando comparado com um eletrocardiógrafo clínico.
Recently, the reliability and cybersecurity aspects of embedded systems for critical domains, such as health and automotive, has increased interest in the research community and awareness to the general public. This thesis addresses the modelling and the implementation of a reliable and secure software architecture for an embedded system aimed at the acquisition and processing of physiological signals, in particular the electrocardiogram (ECG). The work focused CardioWheel, an ongoing project developed by CardioID Technologies, targeting health and automotive applications. These domains demand special requirements for safety and security, and serve as a showcase for the proposed architecture. Accordingly, suitable requirements were first established and the architecture state machine was developed using UML and formally validated using Uppaal and Timed Automata modelling. Then, the threat analysis of the architecture was conducted. Finally, the implementation was realized for an ESP32 microcontroller using the FreeRTOS, the ESP-IDF ecosystem, and specially developed hardware independent layers. The communication layer is based on Bluetooth Low Energy (BLE) and allows the transmission of the data from the end-node to a gateway and finally to the cloud. The system has a Over-The-Air (OTA) component that enables the update of the firmware and the security of this operation was also validated. The proposed architecture was firstly validated using a laboratory prototype. Then, it was deployed to build a small production series of CardioWheel incorporating validation strategies proposed within the context of the ESCEL KDT Valu3s project. Also, a pre-medical trial was conducted at the Hospital de Santa Marta, confirming the reliability and capabilities of our system against a clinical ground-truth.
Recently, the reliability and cybersecurity aspects of embedded systems for critical domains, such as health and automotive, has increased interest in the research community and awareness to the general public. This thesis addresses the modelling and the implementation of a reliable and secure software architecture for an embedded system aimed at the acquisition and processing of physiological signals, in particular the electrocardiogram (ECG). The work focused CardioWheel, an ongoing project developed by CardioID Technologies, targeting health and automotive applications. These domains demand special requirements for safety and security, and serve as a showcase for the proposed architecture. Accordingly, suitable requirements were first established and the architecture state machine was developed using UML and formally validated using Uppaal and Timed Automata modelling. Then, the threat analysis of the architecture was conducted. Finally, the implementation was realized for an ESP32 microcontroller using the FreeRTOS, the ESP-IDF ecosystem, and specially developed hardware independent layers. The communication layer is based on Bluetooth Low Energy (BLE) and allows the transmission of the data from the end-node to a gateway and finally to the cloud. The system has a Over-The-Air (OTA) component that enables the update of the firmware and the security of this operation was also validated. The proposed architecture was firstly validated using a laboratory prototype. Then, it was deployed to build a small production series of CardioWheel incorporating validation strategies proposed within the context of the ESCEL KDT Valu3s project. Also, a pre-medical trial was conducted at the Hospital de Santa Marta, confirming the reliability and capabilities of our system against a clinical ground-truth.
Description
Dissertação para obtenção do Grau de Mestre em Engenharia Informática e de Computadores
Keywords
Sistema crítico Sistema ciberfísico Fiabilidade Cibersegurança Critical system Cyber-physical systems Reliability Cybersecurity
Citation
SANTOS, José Filipe Cruz dos - Reliability and security in wellbeing monitoring embedded systems. Lisboa: Instituto Superior de Engenharia de Lisboa, 2023. Dissertação de Mestrado.