Adaptable optical receiver to recognise different transmitted signal formats

Optical wireless networks have a short-range due to background noise, typically caused by a combination of fluorescent, incandescent, and sun light. Different networks can have different modulation types and different frequencies, which require an intelligent receiver that has the ability to recognise and identify any incoming signal. For the purpose of this thesis, the receiver 'consists of two parts; the optical part and the decision system: The main function of the optical part is to receive. an optical signal and suitably convert it to an electrical signal. Most important is how to improve the low optical signal-to-noise ratio (OSNR) due to the background noise and safety limitations on the optical power. Therefore, the main components of the optical receiver have been discussed and the optimal choices have been suggested. The automatic decision/system has the automatic ability to recognise and identify an incoming signal even if the OSNR is very low. It consists of three identifiers; to differentiate between analogue signals and digital signals, to classify the analogue signals and to recognise common digitally-modulated signals. The analogue identifier may be classified as the statistical approach, and is based on the Hilbert transform and wavelet transform (WT) to extract its key features. Conversely, the digital identifier covers the statistical and decision approaches, and is based on the WT to extract Its key features. .Key features for pattern recognition purposes used were based on their simplicity, speed and their ability to recognise a signal in low SNRs. Therefore, some previously-used features that work in a very low SNR were employed, some features having been extended to include other types of signals, and many new key features were developed and implemented. Despite the fact that the ~ecision system does not require any prior knowledge, the scheme has also been investigated for usefulness in the carrier frequency, SNR and symbol rate estimation which serves to simplify the classification task and help in the demodulation processes.EThOS - Electronic Theses Online ServiceGBUnited Kingdo