Localization of cyborg insects using IMU-based method

This thesis explores the use of Inertial Measurement Unit (IMU) systems to track the movement trajectories of cyborg insects, with a particular focus on IMU-equipped cockroaches. The objective of this research is to accurately localise and track these insects in a controlled indoor environment using...

Full description

Saved in:
Bibliographic Details
Main Author: Fan, Zifu
Other Authors: Hirotaka Sato
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2024
Subjects:
IMU
Online Access:https://hdl.handle.net/10356/177563
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:This thesis explores the use of Inertial Measurement Unit (IMU) systems to track the movement trajectories of cyborg insects, with a particular focus on IMU-equipped cockroaches. The objective of this research is to accurately localise and track these insects in a controlled indoor environment using data-driven machine learning techniques and the VICON motion capture system. In order to improve the reliability of the data, this study employs advanced pre-processing techniques, including wavelet denoising, to refine the IMU data prior to analysis. The core of the study is the development of a random forest regression model that is trained to predict cyborg insect movement based on features extracted from the processed IMU data. The model was validated against data collected by the VICON system, which provides highly accurate spatial localization. The results of the study indicate that while the model provides a reasonably high level of accuracy with respect to standard motion patterns, it encounters limitations with respect to complex movements (especially fast, nonlinear motions). Future work will aim to enhance the model's ability to handle diverse and dynamic motion patterns, and potentially extend it to 3D tracking for more complex environments. This research establishes a foundational approach that combines machine learning with traditional motion tracking techniques and is an important step forward in the automation and accuracy of insect motion analysis.