Multiuser Communication Techniques based on OFDM

Abstract

This thesis deals with the development of a sophisticated multiple access communication technique, Multicarrier Interleave Division Multiple Access (MC-IDMA) and improve- ments in Orthogonal Frequency Division Mulltiplexing-Space Division Multiple Access (OFDM-SDMA) for uplink transmissions. Although different in names both theses tech- niques endeavour to describe a wide range of upcoming communication algorithms and to incorporate three key features that will be common to the next generation communication systems: multiple access capability, resistance to multipath fading and high bandwidth efficiency. The performance of the both systems is demonstrated in the presence of channel impairments. OFDM is almost completely immune to multipath fading effects and IDMA has mul- tiuser capability. MC-IDMA attempts to combine these features, so that we can have Inter Symbol Interference (ISI) free higher data rates for multiple users simultaneously. MC-IDMA uses chip level interleavers as the only means of user separation and de- votes the entire bandwidth expansion to low rate Forward Error Correcting codes (FEC) which provide additional coding gain not present in conventional Code Division Multiple Access (CDMA) or Multicarrier-Code Division Multiple Access (MC-CDMA) systems. An iterative Chip-by-Chip (CBC) multiuser detection algorithm is exploited to sup- press the channel fading and Multiple Access Interference (MAI). To further improve the performance, adaptive subchannel algorithm is used to transmit information only on good conditioned subchannels. To support high order modulations, Multicode MC- IDMA is proposed which can enhance the data rate at the expense of slightly higher power to achieve the same BER. A simplification in a channel estimation algorithm for MC-IDMA system is suggested and the concept of pilot chips is introduced which is a feasible approach because MC-IDMA systems use chip level interleavers and CBC de- tection algorithm. A noval channel estimation algorithm is presented which reduces the joint estimation into several single user estimation problems and increases throughput drastically saving a lot of transmission overhead. The complexity is analyzed and Bit viiError Rate (BER) is computed to depict the performance of MC-IDMA system. The re- sults show that the MC-IDMA system outperforms the conventional MC-CDMA system by supporting a large number of users with enhanced bit error rate performance. Next, a multiple antenna aided, Minimum Bit Error Rate (MBER) based Bell Lab- oratories Layered Space Time (MBER-BLAST) algorithm for OFDM-SDMA communi- cation is presented which overcomes the limitations of detectors when the number of users exceed the number of receiver antennas. A Particle Swarm Optimization (PSO) algorithm is employed for finding the optimum weight vectors for MBER Multiuser De- tector (MUD). PSO is well suited for physically realizable, real-time applications, where low complexity and fast convergence is of absolute importance. Maximum Likelihood (ML) detection algorithm is optimal but it uses exhaustive search and is prohibitively complex. The proposed algorithm outperforms the MBER detector and is capable of achieving performance similar to that attained by ML detector at a significantly lower complexity especially under high user loads. Results show that OFDM-SDMA when used with proposed algorithms promise substantially improved performance and can support throughputs as high as 11 bits/per chip at a BER of 10 −4 . It supports a large number of users by exploiting the capacity advantages of multiple antenna systems in rich scattering environments and offers a good performance-complexity trade-off.