Performance of GMSK and 7/4-DQPSK Modulations in Land Mobile Radio Channels
MetadataShow full item record
Cellular communication systems that provide telephone communication service to mobile users have experienced a rapid growth during the last five years. The current cellular systems are analog in nature. These systems employ frequency modulation. As the spectrum gets crowded, bandwidth efficient digital systems will find increasing application. Gaussian Minimum Shift Keying (GMSK) and 7r/4-Differential Quadrature Phase Shift Keying (7r/ 4-DQPSK) have been identified as the modulation methods for the European and the North American digital cellular systems respectively. The communication channel for the cellular systems is commonly referred to as the land mobile radio channel. The land mobile radio channel can be modeled as a frequency selective Rayleigh fading channel corrupted by additive white Gaussian noise. In addition to cochannel interference, delayed signals and Doppler effects have to be considered. In this thesis the effect of predetection filter bandwidth on the bit error rate performance of GMSK with one-bit differential detection in this environment has been analyzed. The bit error rate performance of GMSK with two-bit differential detection has also been analyzed. Numerical computation is used to obtain GMSK bit error rate results for various combinations of channel parameters. The computed bit error rate results for one-bit differential detection indicate that the normalized predetection filter bandwidth can be kept at one for all types of land mobile radio channels. The computed bit error rate results for two-bit differential detection can be used to predict the performance of GMSK more accurately in land mobile radio channels. An upper bound on the bit error rate is also derived for convolutional coded 7r/ 4-DQPSK modulation with frequency-offset diversity and soft decision Viterbi decoding. Based on the numerically computed bit error rate bound it was found that the combination of frequency- offset diversity and convolutional coding results in a very significant improvement compared to a simple differential detector receiver which does not have any diversity and soft decision Viterbi decoding. The bound derived is verified by simulation to get an estimate of the bit error rate. It was found that the bound is quite tight and can be used to predict the performance of convolutional coded 7r/4-DQPSK.