Recently, multi-carrier transmission in the form of orthogonal frequency division multiplexing (OFDM) attracts a lot of interest because of its ability to support high-rate data transmission in wireless applications including digital video broadcast (DVB), digital audio broadcast (DAB) and wireless Internet access. Many literatures have demonstrated that multi-carrier transmission achieves optimal performance by varying the allocation of the data rates and the transmit power among sub-channels when the information of the channel transfer function is available....[ Read more ]

Recently, multi-carrier transmission in the form of orthogonal frequency division multiplexing (OFDM) attracts a lot of interest because of its ability to support high-rate data transmission in wireless applications including digital video broadcast (DVB), digital audio broadcast (DAB) and wireless Internet access. Many literatures have demonstrated that multi-carrier transmission achieves optimal performance by varying the allocation of the data rates and the transmit power among sub-channels when the information of the channel transfer function is available.

In this thesis, we study the optimization problem of minimizing the overall transmit power for a given data rate, a given bandwidth and a fixed bit error rate (BER) in OFDM transmission. We first develop a low complexity optimal bit and power allocation algorithm for optimum power distribution over sub-channels. Then, we design an adaptive trellis coded modulation (ATCM) scheme using a rate-l/2 convolutional encoder and M-QAM signal constellations to support the application of adaptive modulation in each sub-channel. Using the developed algorithm and designed ATCM scheme, we evaluate the performance of the adaptive trellis coded OFDM system under the assumption of perfect channel information in the transmitter and compare it with that of the OFDM system based on fixed trellis coded modulation.

In this thesis, we also develop an adaptive tracking algorithm to adaptively update the bit and power allocation for OFDM transmission to accommodate the time variations of the channel. This adaptive tracking approach utilizes the previous allocation for initialization, and then iteratively refine the allocation until it achieves the optimal bit and power allocation. We prove a theorem which guarantees the convergence of the algorithm to the optimal solution. Simulation results show that the adaptive tracking algorithm converges rapidly to the optimal solution and can be implemented efficiently.

In practical implementations, the assumption of perfect channel information in the transmitter may not be achieved due to channel estimation errors or the time-varying nature of the channel. Therefore, we study the impact and effects of imperfect channel knowledge on the performance of the adaptive trellis coded OFDM system.