Research On Quantization Processing And Realization Of Linear Dispersion Codes

MIMO technology, namely Multiple Input Multiple Output technology, is recently regarded as one of the most effective technology to improve the performance of communication systems and increase the channel capacity. So MIMO is regarded as the most potential technology for the third generation wireless communication system and even the fourth generation wireless communication system.The realization of MIMO technology depends on multiple antenna technology, and it acts in the way of Space-Time Coding. Space-Time Coding is a signal transmission scheme that maps data symbols to the two dimensions of space and time and then sends them with multiple antennas. Space-Time codes are generated according to some optimal schemes. Among the space-time codes which are researched much now are STBC (Space Time Block Codes), BLAST (Bell Labs Layered Space-Time), and LDCs (Linear Dispersion Codes). These codes all have their advantages and disadvantages.Among the above three Space-Time Codes, LDCs received much attention recently, this is because on the one hand, LDCs subsume VBLAST and many STBC codes as special cases and LDCs usually have better performance, on the other hand, LDCs can be encoded and decoded conveniently: the encoding can be performed only by some linear combination operations and as for the decoding, any decoding algorithms that can be used to decode VBLAST can be taken advantage to decode LDCs. Moreover, LDCs can be used in any systems that have any number of transmit and receive antennas.LDCs are generated according to some generation rules, and the rules are always include some optimal problems, so the codes that can fulfill those rules usually have infinite precision. However, the codes that can be applied to practice usually have finite precision, then, the problem of how accurate the codes should be expressed arises. The thesis researched the decision of precision of LDCs. Besides, how to decode LDCs in practical systems is also an important issue. A LDCs system is equivalent to a VBLAST system, so any decoding schemes that can be used to decode VBLAST can also be applied to LDCs decoding. This thesis realizes the LDCs decoding algorithm based on QR-decomposition making use of FPGA. This realization depends on systolic arrays. The results demonstrate that the throughput is advanced at the cost of more resource consumption.The thesis first introduces the research background of wireless communication and the research situation of MIMO and Space-Time technology. Then it talks about some useful knowledge related to MIMO and Space-Time technology, such as system model, channel capacity, the fundamental tradeoff of diversity and multiplexing, the basic rules of Space-Time designs, codes design of LDCs, etc. Next, the thesis gives a general overview of some commonly-used space-time codes. Then, the dissertation proposes the concept of Quantized Linear Dispersion Codes (QLDCs). Through quantizing the real part and image part of each element in the code matrices, we tries to find approximate codes that consume little memory resources, but have almost the same Mutual Information(MI) and Bit Error Rate performance as the original codes. Simulation results show that such approximate codes can be generated through properly quantizing the original LDCs codes. ThenAt last, we realize the QLDCs decoding algorithm based on QR-decomposition using Altera FPGA(Field Programmable Gate Array). Results show that QR-decomposition can get good BER performance and is easy to realize using FPGA.