PERFORMANCE ANALYSIS OF CROSS QAM OVER FADING CHANNELS WITH MULTIPLE DIVERSITY COMBINING

Abstract

QAM has been widely used as a bandwidth efficient digital modulation scheme. As it makes use both amplitude and phase levels for symbols transmition. Initially QAM was known for its square shaped structure (square QAM) that had even number of bits per symbol. System’s limitations regarding transmitted power created need for another version of QAM that can have odd number of bits per symbol (rectangular QAM). Later, Smith proposed an effective type of QAM (cross QAM) that not only had odd number of bits per symbol but also it was an energy efficient scheme as compared to the rectangular QAM. Implementation complexity of both modulation schemes were quite similar but the advantage that cross QAM had over rectangular QAM was better performance. In this dissertation, cross QAM will be analyzed over AWGN and different fading channels including Rayleigh, Nakagami-m, Nakagami-q, Nakagami-n and Beckmann. Later, SEP performance of cross QAM is optimized by utilizing multiple diversity combining techniques including selection combining (SC), maximal ratio combining (MRC) and generalized selection combining (GSC). Also SEP performance of cross and rectangular QAM has been compared for Equal Energy Case (EEC) over all the previously described fading channels. Cross QAM has wide range of applications such as adaptive modulation scheme, where the constellation size is increased according to the channel quality. When channel quality is good, the constellation size is increased by ‘k+1’ bits per symbol. If we are to consider only even bits per symbol then increment size would be ‘k+2’ bits per symbol (we have to go from 16 to 64 to 256 QAM...). XQAM however provides the flexibility to reduce the intermediate step size from ‘k+2’ to ‘k+1’ bits per symbol (we need to go from 16 to 32 to 64 QAM…). Use of XQAM for single bit increase makes the change relatively smoother and enables the system to perform better over a required data rate. cross QAM with symbol length of 5 (M=32) to 15 (M=32786) bits are commonly used in ADSL and VDSL for high data rate applications, also 32 and 128 XQAM had been applied in digital video broadcasting. XQAM also has many applications regarding blind equalization, where the channel’s impulse response is estimated by the equalizer without prior knowledge of channel behavior.