Differential Modulation in LTE-Downlink

This diploma thesis evaluates the performance of differential modulation applied to the 3GPP Long Term Evolution downlink. The focus specifically lies on single input single output transmissions with only one user. The Long Term Evolution downlink is based on channel estimation. The reference symbols, required for channel estimation cannot be used to transmit data, which reduces the spectral efficiency. In order to increase the spectral efficiency, in this thesis noncoherent detection is considered. With noncoherent detection no channel knowledge is needed. Thus, the reference symbols for channel estimation are no longer required. Two basic modulation schemes, namely the frequency first and the time first modulation scheme, are presented. Their performance is investigated both for frequency selective linear time-invariant and linear time-variant channels and compared to coherent detection with least squares channel estimation and zero forcing equalization. As performance criteria the bit error ratio and the throughput over the signal to noise ratio are considered. In the investigations it is discovered that coherent detection outperforms conventional noncoherent detection. To improve the efficiency also multiple symbol differential detection is applied. Furthermore, the correct selection of the modulation scheme for different channels is discussed.