Comparison of distributed space and frequency interference alignment

This paper addresses multiple access in MIMO wireless networks. It compares two distributed interference alignment techniques, in space and in frequency, that both aim at removing interferences through orthogonalization. The theoretical advantages and drawbacks of each technique are highlighted. Then, the multiplexing gain and the computational complexity are analytically evaluated. It is shown that frequency interference alignment achieves a higher multiplexing gain than space interference alignment in realistic network conditions, and always requires a lower computational complexity. Finally, the performances of both algorithms with varying number of interfering links are assessed via numerical simulations. Space interference alignment fulfills complete interference suppression, and consequently provides the same spectral efficiency to all links, regardless of the number of interfering links. Frequency interference alignment is less efficient in terms of INR, and the spectral efficiency per link thus decreases when the number of links increases. Nevertheless, the spectral efficiency is always higher with frequency interference alignment than with space interference alignment.