|
The vision of future generation of wireless systems will be the provision of broadband access, seamless global roaming, Internet/data/voice and highly interactive multimedia services such as multiparty videoconference and virtual telepresence anywhere anytime. Providing such ubiquitous high-capacity high-quality transmission will require several Mbps to be achieved over air. Though current 2.5G and evolving 3G technologies promise burst rates up to 348kbps and 2Mbps, respectively, the average throughput per user is not expected to be more than 171kbps during busy hours. To rectify this, various advanced techniques have been proposed, among which multi-element transmitter and receiver antenna configurations known as multiple-input multiple-output (MIMO) antenna systems have recently emerged as a significant breakthrough for enhancing the capacity far beyond systems in use today. MIMO techniques have already been proposed as an option for 3GPP (Third Generation Partnership Project) Release 6, and the standardisation of a well-known MIMO scheme called BLAST (Bell Labs Layered Space-Time) architecture is underway.Physically, radio signals, transmitted from or received at, different antennas (if they are separated far apart) are very different that they can be distinguished from their so-called spatial signatures. This principle allows multiple signal streams to be multiplexed in spatial domain, leading to tremendous capacity achievable by MIMO antenna. The technical challenge, however, is that the optimum MIMO performance requires maximum-likelihood (ML) decoding, which is unfortunately non-deterministic polynomial-time hard (NP-hard). Neither a compact mobile device nor a powerful base station is feasible even if the number of multiplexed signals is moderate (e.g., 5). Worse of all, in real environments, the presence of co-channel interference (other signals that share with the same radio channel) as a result of frequency reuse for cellular systems, will further cause ML decoding even more difficult to implement.The objective of this project is to design low-complexity MIMO receivers for high-performance decoding and study the use of such receivers in multiuser environments in both up (from many mobile stations to a base station) and downlinks (from a base station to many mobile stations) with appropriate transmitter designs.Suboptimal decoding leading to reduced complexity has long been investigated for MIMO detection. Though relatively practical, it either works only for specific modulation or needs to scarify much diversity that already exists in the channel. Besides, most suboptimal detectors are designed for single MIMO link only and require the number of receive antennas to be at least equal to the number of multiplexed signal streams. It is, however, not clear how they would work in a cellular network where the same frequency channel is more aggressively reused and the antenna array is largely overloaded. Such scenario will be dealt with in this project. The study is important, for ultimately spectral-efficient cellular networks ought to allow every radio channel to be shared by as many users as possible, which gives rise to overloaded antenna array. In this respect, we aim to develop a non-linear signal processing scheme that can place more nulls than the number of receive antennas, exceeding the limit of what linear processing can achieve.Subject to different system constraints for single, or multiuser, uplink or downlink channels, various MIMO receivers with the best tradeoff between complexity and performance will be devised that makes MIMO technologies practically viable. The developed detectors should be scalable for a system with a large number of users/antennas, and as a consequence, MIMO technologies such as BLAST and space-time coding (in 3G/4G+ systems) can be practically deployed to yield promising performance in interference-limited cellular environments.
|