Banca de DEFESA: PEDRO HENRIQUE MEIRA DE ANDRADE

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : PEDRO HENRIQUE MEIRA DE ANDRADE
DATE: 30/08/2024
TIME: 08:00
LOCAL: Remoto
TITLE:

A TinyML Incremental Learning Approach for Outlier Processing and Forecasting

KEY WORDS:

Internet of Things, Edge Computing, TinyML, Tiny Machine Learning, Incremental Learning, Continual Learning, Outlier Detection, Outlier Correction, Microcontrollers.

PAGES: 100
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUBÁREA: Eletrônica Industrial, Sistemas e Controles Eletrônicos
SPECIALTY: Automação Eletrônica de Processos Elétricos e Industriais
SUMMARY:

The Internet of Things (IoT) is a paradigm where computing and connectivity capabilities are embedded into objects, connecting them to the Internet. Acknowledged as a crucial and emerging technological area, IoT holds significant potential to enhance quality of life, optimize industrial processes, and offer more applications to everyday objects. With the increasing number of IoT-connected devices, there arises a need for infrastructure to manage the vast volume of generated data. In this context, Edge Computing stands out by processing data close to its source, leaving only heavier processing tasks for central servers. Edge processing enables the development of optimized machine learning algorithms, known as Tiny Machine Learning (TinyML). By employing lightweight and optimized algorithms, TinyML offers advantages such as reduced latency, improved energy efficiency, and increased autonomy for devices operating in remote or isolated applications. In the field of TinyML, implementing machine learning techniques on resource-constrained devices like microcontrollers poses significant challenges, including outlier detection and correction. This work addresses the outlier problem, crucial for both academic research and industrial applications in domains such as energy measurements, health data, industrial systems, and automotive applications. Three algorithms were developed: TEDA-RLS and TEDA-Forecasting for outlier processing, focusing on detection and correction, and TEDA-Ensemble for forecasting. The developed algorithms are based on the concept of Incremental Learning, as they have the capability to continuously learn from new data inputs. The proposed algorithms were compared with established techniques in the literature such as XGBoosting, Long Short-Term Memory (LSTM), Linear Regression, and k-Nearest Neighbors (KNN), yielding promising results with low error rates and minimal energy consumption. Finally, the outlier processing algorithms were successfully embedded in a microcontroller, confirming the feasibility of the TinyML approach.

COMMITTEE MEMBERS:
Presidente - 2885532 - IVANOVITCH MEDEIROS DANTAS DA SILVA
Interno - 1153006 - LUIZ AFFONSO HENDERSON GUEDES DE OLIVEIRA
Externo ao Programa - 3216921 - TIAGO TAVARES LEITE BARROS - UFRNExterno à Instituição - DANIEL GOUVEIA COSTA - FEUP
Externo à Instituição - JUAN MOISES MAURICIO VILLANUEVA - UFPB