Single-channel source separation of communication signals
Date of Issue2016
School of Electrical and Electronic Engineering
Single-channel signal separation has received more and more attention in recent years due to the improved capacity it offers. In this project, two improved Phase Shift Keying (PSK) symbol detectors are developed for separation of signals with single modulation or mixed modulation, while the existing PSK detector can only separate signals with single modulation. The Symbol Error Rate (SER) performance and the running time of these two detectors are validated through simulations in Matlab. The algorithm combines the initialization and optimization technique with gradient descent strategy. The first proposed detector is the Semidefinite Programming (SDP) +PSK detector, which is based on the SDP initialization method. The second proposed detector is the Random initialization+PSK detector which combines the random initialized signal with the noise level constraint. The proposed detectors are first tested for separation of two single-modulation signals and compared with the existing PSK detector. They are verified for separation of two signals with different modulation afterwards. Lastly, they are applied in the real channel simulation for future validation. The low-SER performance of the two proposed improved detectors have been validated at the end and the main object of this project, which is developing improved PSK detectors for mixed-modulation signal separation, has been achieved. From the perspective of the running time, the second proposed detector is recommended for further study with arbitrary modulation in a more complicated channel.
Final Year Project (FYP)
Nanyang Technological University