Skip to Main Content
Wavelets can be used to transform a signal into a multi-dimensional signal where each dimension represents one wavelet resolution, such that a machine learning classifier, such as artificial neural network (ANN), may be used to then classify the received signal and recover the transmitted information. Since there is no upper limit for wavelet resolution and wavelet resolution produces highly redundant coefficients, computational difficulties are when signal need to be classified. Here we demonstrate the use of dimension reduction techniques to visualise indoor optical wireless communication (OWC) signal in the presence of artificial light interference, scale reduction technique to for efficient classification and the resulting decoding errors.
Date of Conference: 21-23 July 2010