| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.35 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Nesta dissertação, abordamos um desafio crucial no âmbito das Redes de Sensores Sem
Fios (RSSFs) baseadas no padrão IEEE 802.15.4: a eficiência energética, particularmente
relevante diante do crescente interesse na expansão da Internet das Coisas (IoT). Em um
cenário em que cidades inteligentes, agricultura de precisão e aplicações de saúde remota
impulsionam a evolução da IoT, a eficiência energética dos nós sensores emerge como
uma consideração crítica.
Nossa análise concentra-se nos protocolos de acesso ao meio existentes, com especial
ênfase no Carrier Sense Multiple Access (CSMA), protocolo tradicionalmente adotado
neste padrão. Em contrapartida, propomos a adoção e otimização do protocolo Slotted
ALOHA como alternativa, visando superar as limitações identificadas nos protocolos
convencionais e alinhando-se com os requisitos da IoT. Apresentamos variantes
aprimoradas do Slotted ALOHA, incorporando técnicas como Binary Exponential
Backoff (BEB) e aprendizado de reforço, nomeadamente o Q-learning. A implementação
dessas melhorias visa não apenas otimizar o rendimento da rede, mas também aprimorar
a eficiência energética, reduzir o atraso nas transmissões e proporcionar benefícios
substanciais em diversas métricas operacionais, fundamentais para o sucesso da IoT.
Através de simulações no MATLAB, demonstramos que o Slotted ALOHA aprimorado
com Q-learning oferece benefícios significativos em termos de consumo energético e
redução do atraso, destacando-se como uma alternativa promissora ao CSMA para
RSSFs na era da IoT. Essa contribuição não só avança na compreensão das dinâmicas de
eficiência energética, mas também propõe soluções práticas e inovadoras para otimizar
o desempenho global das RSSFs em ambientes diversos e desafiadores.
In this dissertation, we address a crucial challenge in the context of Wireless Sensor Networks (WSNs) based on the IEEE 802.15.4 standard: energy efficiency, particularly relevant given the growing interest in the expansion of the Internet of Things (IoT). In a scenario where smart cities, precision agriculture, and remote health applications drive the evolution of IoT, the energy efficiency of sensor nodes emerges as a critical consideration. Our analysis focuses on existing medium access protocols, emphasizing Carrier Sense Multiple Access (CSMA), a protocol traditionally adopted in this standard. Conversely, we propose adopting and optimizing the Slotted ALOHA protocol as an alternative to overcome identified limitations in conventional protocols and align with IoT requirements. We present enhanced variants of Slotted ALOHA, incorporating techniques such as Binary Exponential Backoff (BEB) and reinforcement learning, notably Q-learning. The implementation of these improvements aims not only to optimize network performance but also to enhance energy efficiency, reduce transmission delay, and provide substantial benefits in various operational metrics, crucial for IoT success. Through simulations in MATLAB, we demonstrate that Slotted ALOHA enhanced with Q-learning offers significant benefits in energy consumption reduction and delay reduction. It is a promising alternative to CSMA for WSNs in the IoT era. This contribution advances the understanding of energy efficiency dynamics and proposes practical and innovative solutions to optimize the overall performance of WSNs in diverse and challenging environments.
In this dissertation, we address a crucial challenge in the context of Wireless Sensor Networks (WSNs) based on the IEEE 802.15.4 standard: energy efficiency, particularly relevant given the growing interest in the expansion of the Internet of Things (IoT). In a scenario where smart cities, precision agriculture, and remote health applications drive the evolution of IoT, the energy efficiency of sensor nodes emerges as a critical consideration. Our analysis focuses on existing medium access protocols, emphasizing Carrier Sense Multiple Access (CSMA), a protocol traditionally adopted in this standard. Conversely, we propose adopting and optimizing the Slotted ALOHA protocol as an alternative to overcome identified limitations in conventional protocols and align with IoT requirements. We present enhanced variants of Slotted ALOHA, incorporating techniques such as Binary Exponential Backoff (BEB) and reinforcement learning, notably Q-learning. The implementation of these improvements aims not only to optimize network performance but also to enhance energy efficiency, reduce transmission delay, and provide substantial benefits in various operational metrics, crucial for IoT success. Through simulations in MATLAB, we demonstrate that Slotted ALOHA enhanced with Q-learning offers significant benefits in energy consumption reduction and delay reduction. It is a promising alternative to CSMA for WSNs in the IoT era. This contribution advances the understanding of energy efficiency dynamics and proposes practical and innovative solutions to optimize the overall performance of WSNs in diverse and challenging environments.
Description
Keywords
Binary Exponential Backoff (Beb) Carrier Sense Multiple Access (Csma) Eficiência Energética IEEE 802.15.4 Internet das Coisas (Iot) Q-Learning Redes de Sensores Sem Fios (Rssfs) Slottedaloha