A simplified approach to multi-carrier modulation
There is a significant demand for a decrease in the size, weight and power (SWaP) associated with wireless systems. In recent years, multiple-input, multiple-output (MIMO) wireless systems have received considerable attention due to the high data rates they provide. Orthogonal frequency division multiplexing (OFDM), a digital multi-carrier modulation technique, is well suited to be used in MIMO systems as it provides the ability to operate in frequency-selective channel environments. When OFDM is combined with the capacity increase provided by MIMO systems, the result is a very successful communication system. In this research, a reduced-complexity MIMO OFDM system is advanced. The proposed system is multiplier-less and thus requires a simpler digital hardware implementation. As a result, the chip area, power consumption and cost associated with the MIMO OFDM system can be significantly reduced.^ The reduction in complexity is obtained via modification to conventional Fast Fourier Transform (FFT) and Inverse Fast Fourier Transform (IFFT) algorithms necessary to implement OFDM multi-carrier modulation. System computational complexity is reduced by quantizing what are known as “twiddle factors” in traditional FFT algorithms such as the Radix-2 and Radix-4. The quantization allows for all multiplications to be done with a value of one, negative one, zero or a power of two. Ensuring that all multiplications are performed with any of the aforementioned values results in a transform where all multiplications are considered trivial. Replacing standard multiplications with trivial multiplications significantly reduces system computational complexity. As an example, the complexity associated with the implementation of the rounded FFT as compared to a conventional Radix-4 FFT is reduced by 47% when numerical values are represented with 16 bits. Depending on the application, different quantization levels can be utilized in order to obtain the necessary performance characteristics. As the number of quantization levels grows, the system capability increasingly approaches the performance of a system that uses the conventional transforms. When applied to MIMO OFDM systems, the computational savings are significant as the combination of the IFFT and FFT algorithms are implemented for every spatial stream (i.e. antenna). As such, the simplified approach provides a system that is a lower-cost, practical alternative to the MIMO OFDM systems used today.^
Steven Walter, Purdue University.
Engineering, Electronics and Electrical