| Title |
Dynamic Observation of ML for Robotic Perception: A Survey of Recent Innovations |
| Abstract |
This project outlines the development and deployment of a non-contact vibration sensor designed to
gather data from rotating machinery to facilitate early detection of bearing faults. The Hilbert
transform is employed to reduce noise in the vibration signals, which are then processed using Principal
Component Analysis (PCA) for dimensionality reduction and Sequential Floating Forward Selection
(SFFS) for feature selection. Key features are utilized to identify and classify various bearing problems
through Support Vector Machines (SVM) and Artificial Neural Network (ANN) algorithms. This
approach offers a proactive and efficient solution for monitoring bearing health, emphasizing rapid
fault detection and resulting in considerable savings in time, effort, and maintenance costs. |
| Keywords |
Machine Learning, Fault Prediction, Fuzzy Convolution Neural Network (FCNN), Heterogeneous Sensing Data Fusion |
| Reserch Area |
Engineering |
| Reserch Paper |
AIJFR2404002 - V2 I4 - 13-20.pdf |
| Author(s) |
J. Bhaskar, L. M. Bharathi |
| Country |
India |
