Novel Insect Detection using Low-Cost, Field-Deployable, Stand-Alone Microcontrollers
Faculty Mentor Name
Seth McNeill
Format Preference
Poster
Abstract
The goal of the project is using machine learning algorithms on embedded system to detect harmful insect calls, specifically focusing on psyllid calls, in collaboration with USDA researchers. Additionally, the project will help working towards a redesign of the robots for Dr. McNeill’s microcontrollers course, with the objective of providing each student with their own robot rather than solely relying on the robots in the lab. The ESP32 and RP2040 development platforms has shown the feasibility of employing more cost-effective and user-friendly processors, along with a VS Code development environment that integrates git usage into the class. Dr. McNeill has expressed a desire to incorporate git into the class, but the current development environment presents challenges in doing so. Moreover, this research aims to build a library of working materials that Dr. McNeill can utilize in teaching an Embedded Machine Learning course. Following the achievement of these fundamental goals, we plan to start researching ways of detecting illegal drones crossing the USA/Mexico border, using cheap microcontrollers for the acoustic side of detection, which aligns with our ongoing development efforts. We also aim to use connections from the industrial sector, where there is interest in using embedded machine learning to monitor machine health. In summary, this project seeks to address research, industrial, and educational objectives.
Novel Insect Detection using Low-Cost, Field-Deployable, Stand-Alone Microcontrollers
The goal of the project is using machine learning algorithms on embedded system to detect harmful insect calls, specifically focusing on psyllid calls, in collaboration with USDA researchers. Additionally, the project will help working towards a redesign of the robots for Dr. McNeill’s microcontrollers course, with the objective of providing each student with their own robot rather than solely relying on the robots in the lab. The ESP32 and RP2040 development platforms has shown the feasibility of employing more cost-effective and user-friendly processors, along with a VS Code development environment that integrates git usage into the class. Dr. McNeill has expressed a desire to incorporate git into the class, but the current development environment presents challenges in doing so. Moreover, this research aims to build a library of working materials that Dr. McNeill can utilize in teaching an Embedded Machine Learning course. Following the achievement of these fundamental goals, we plan to start researching ways of detecting illegal drones crossing the USA/Mexico border, using cheap microcontrollers for the acoustic side of detection, which aligns with our ongoing development efforts. We also aim to use connections from the industrial sector, where there is interest in using embedded machine learning to monitor machine health. In summary, this project seeks to address research, industrial, and educational objectives.