#TITLE_ALTERNATIVE#

• Main Author: Andri Wijaya (NIM 132 03 147), Michael
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/8591
• Institution: Institut Teknologi Bandung

Abstract: <br /> <br /> <br /> <br /> <br /> A simple diversity system to overcome the effect of wireless channel fading is introduced by Alamouti. Alamouti system works under the assumption that the system perfectly knows the conditions of the wireless channel. This assumption is further complemented by Alamouti with the no channel estimation channel estimation technique for the channel estimator part in the system. <br /> <br /> <br /> <br /> <br /> In this final project, Alamouti system is evaluated in four various channel estimation: perfect channel knowledge assumption, no channel estimation technique, Least Mean Square (LMS) technique, and Neural Network (NN) based technique: Time Delay Neural Network (TDNN) technique. Alamouti system with perfect channel knowledge assumption and no channel estimation technique is simulated at baseband level in the form of bit stream. The other channel estimation variation (LMS and TDNN) is simulated as bit stream which went through the oversampling process. Then, the best channel detection tehnique is decided, judging both the performance and complexity, according to the simulated signal type. <br /> <br /> <br /> <br /> <br /> From MATLAB simulation, we can get a BER (Bit Error Rate) curve for each channel estimation that is better from BER curve for BPSK modulated no-diversity system (SISO) in flat fading channel. Channel estimation with perfect channel knowledge assumption acquires the curve at transfer rate above 1 bps. No channel estimation technique acquires the curve at transfer rate above 1 kbps. LMS channel estimation technique acquires the curve at transfer rate above 5.3 Mbps, while TDNN at above 4.6 Mbps. <br />