The Design and Implementation of a High Efficiency and Low Power Qr Decomposition Architecture in Mimo Detection Systems

MIMO-OFDM technology has been widely adopted in the latest and next generation broadband wireless communication standards, including IEEE 802.11n, IEEE 802.16e-2005 (mobile WiMAX), 3GPP-LTE, and LTE-Advanced. QR-Decomposition mechanism, which pre-processes the channel matrix for MIMO detector to reduce its design complexity, is an essential and fundamental component of all MIMO receivers. As the data rates become higher and higher and the portability of the mobile wireless broadband communication devices is more desired, having a high efficiency QR-Decomposition (QRD) algorithm and a low power implementation of it is crucial and challenge. In this project, we will design and develop a novel CORDIC-based QRD architecture to reduce its clock latency and hardware complexity and hence with higher computational efficiency comparing to the other latest CORDIC-based QRD architectures. We propose a hybrid design which will decompose the complex channel matrix directly without matrix expansion in multi-dimensional manners. We also modified the basic CORDIC operations to allow the rotation operations to be executed in parallel with vectoring operations and hence it performs high-parallelized annihilations and rotations of greater than two dimensional column vectors. Moreover, we will also adopt the modified scale factor correction scheme to avoid the need of complex multipliers for the scaling factor correction. With all these together, we can have a low latency and high performance QRD architecture. We will also design new algorithmic noise-tolerance techniques for the core logics in our VLSI implementation for low power implementation to scale down the supply voltage while maintaining acceptable BER for the MIMO detection systems. We will verify it by Taiwan Semiconductor Manufacturing Company 90n CMOS technology through the fabrication services from National Chip Implementation Center of Taiwan. These are the two main goals in this project to achieve and the major contribution of this proposed research is to deliver an IP design of a high efficient and low power QRD for MIMO receiver systems.

Project ID:PB10007-7224
External Project ID:NSC100-2221-E182-039