UTILIZATION OF CHANNEL STATE INFORMATION FEEDBACK FOR INCREASING CAPACITY AND QUALITY ON OFDMA RADIO RESOURCES ALLOCATION

• Main Author: Prasetya, Budi
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/53184
• Institution: Institut Teknologi Bandung

In studies that have been conducted on the radio resources allocation of the Orthogonal Frequency Division Multiple Access (OFDMA) system, almost all of them contribute to the algorithm for increasing spectral efficiency, throughput, fairness with lower complexity. As an input algorithm, used the channel state information at the transmitter in the form of channel gain only and has not discussed about improving the communication performance. There has been research to improve it on multicarrier systems by rotate modulation process but it is still open-loop. In this study, we propose a new method for utilizing feedback on channel conditions in the form of a complex value with the aim of increasing the capacity and quality of the OFDMA radio resource allocation systems. To increase capacity, we propose a chunk forming process, namely the grouping of several subcarriers into one chunk using the K-Means Clustering algorithm, this algorithm was chosen because of the simplicity of the computation process. To improve quality, we propose a new method, which is to combine power loading and phase shift simultaneously. Power loading and phase shift are generally applied separately to improve the quality of digital communication on multicarrier systems. Our propose method uses the principle of channel equalization, but we apply it to the transmitter. The simulation results of the chunk forming process in the resource allocation scheme can increase throughput when compared to conventional chunk forming, and show that the smaller SSE (Sum Squared Error) value of the clustering process, the greater the SSB (Sum of Squared Between groups). The simulation results also show that the combined model of power loading and phase shift gives a very significant probability of bit error improvement when compared to the application of power loading alone or phase shifting alone. In an ideal feedback condition, the form of the received signal is the same as the transmitted signal, so that the probability of bit error produced is the same as the system that passes through the AWGN channel only. iv In real conditions, the process of transmitting the feedback signal and the channel estimation process in the user results the feedback delay and error. The results of tests on all types of modulation show the sistem performance with the feedback delay of 10 samples has not been degraded. The results of tests on low-order modulation such as BPSK and QPSK, show that the probability of bit error the system does not experience significant degradation with the delay. However, for higher order modulation such as 16QAM and 64QAM, the feedback delay causes the probability of bit error curve to degrade. The results of testing the error feedback effect, show the probability of bit error curve changes very crucially. For the feedback symbol error rate value of 0.001, the system's probability of bit error has degraded for all modulation orders. Thus, the effect of error on feedback is more dominant than the effect of delay on feedback. Feedback on channel conditions that are used to improve the quality of communication sometimes experiences deep fade conditions, therefore we also analyze the addition of a limiter to the feedback to anticipate if a deep fade occurs. With the simulations, we display the visualization of the signal in each process on the transmitter and receiver sides for different levels of limiter level. The test results show if we choose a higher limiter level, then we get the worse the probability of bit error. In low order modulation, the result of probability of bit error is still good even though the limiter level process is added. Limiter level that is still safe for all modulation orders is 0.01.