A mode identification system for a reconfigurable terminal using Time Frequency analysis and a non-parametric classifier

The use of Time Frequency (TF) analysis is proposed as signal processing technique combined with a pattern recognition approach, for identifying the transmission modes in indoor wireless environment with a reconfigurable mobile terminal based on Software Radio techniques. In particular, a Software Radio device is considered aiming at the identification of the presence of two co-existent communication modes as Bluetooth, based on Frequency Hopping - Code Division Multiple Access (FH-CDMA), and IEEE WLAN 802.11b, based on Direct Sequence - Code Division Multiple Access (DS-CDMA). A pattern recognition approach will be presented, where TF analysis is employed for feature extraction, and a multi hypotheses k nearest neighbors (k-NN) non parametric classifier is used. Results in terms of error classification probability, expressed as relative error frequency, will be provided.