Banca de QUALIFICAÇÃO: HAGI JAKOBSON DANTAS DA COSTA

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : HAGI JAKOBSON DANTAS DA COSTA
DATE: 28/08/2024
TIME: 14:00
LOCAL: Remoto
TITLE:

An Evolving Multivariate Times Series Compression Algorithm for IoT Applications

KEY WORDS:

IoT, TinyML, evolving Algorithm, data compression, OBD-II Edge

PAGES: 45
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUMMARY:

The Internet of Things (IoT) is revolutionizing the way devices interact and share data, especially in applications such as vehicle monitoring. However, transmitting large volumes of data in real-time can result in high latency and significant energy consumption. In this context, Tiny Machine Learning (TinyML) emerges as a promising solution, enabling the execution of machine learning models on embedded devices with limited resources. This work aims to develop an online multidimensional compression algorithm specifically for TinyML, using Typicality and Eccentricity Data Analytics (TEDA). The algorithm is based on the eccentricity of the data and does not require pre-established mathematical models or any assumptions about data distribution, contributing to minimizing latency and energy consumption during data transmission. The methodology involves applying the algorithm in a case study with the OBD-II Freematics device, focused on vehicle monitoring. Preliminary results indicate that the proposed algorithm offers a significant improvement in terms of execution time and accuracy compared to traditional compression methods. These results highlight the algorithm's potential to optimize the performance of embedded IoT systems, contributing to the efficiency and sustainability of vehicular applications.

COMMITTEE MEMBERS:
Presidente - 2885532 - IVANOVITCH MEDEIROS DANTAS DA SILVA
Externo à Instituição - IGNACIO SANCHES GENDRIZ
Externa à Instituição - MARIANNE BATISTA DINIZ DA SILVA - UFAL