This paper first discusses the maximum likelihood (ML) classifier for automatic classification of digital modulations. The classifier is optimum for classification of phase-amplitude modulated signals under ideal environment. However, this is not the case in the presence of phase offset owing to inaccurate estimation. In this paper, we propose a novel non-coherent ML classifier to mitigate the effect phase offset. The non-coherent ML classifier adopts a pre-classification phase correction stage through a closed form estimator based on Higher Order Statistics. Experimental results show improvement of classification accuracy at reasonable signal to noise ratio.
|Title of host publication||18th International Conference on Pattern Recognition|
|Number of pages||4|
|Publication status||Published - 2006|
ASJC Scopus subject areas
- Electrical and Electronic Engineering
- Computer Vision and Pattern Recognition
- Hardware and Architecture