The performance of successive interference cancellation in random wireless networks

This paper provides a unified framework to study the performance gain of successive interference cancellation (SIC) in d-dimensional interference-limited networks with arbitrary fading distribution and power-law path loss. We derive bounds on the mean number of users that can be successively decoded and the probability of successively decoding k users. Our results suggest that, without power control, the marginal benefit of enabling the receiver to successively decode k users diminishes very fast with k, especially in networks of high dimensions and small path loss exponent. On the other hand, SIC is more beneficial when the users are clustered around the receiver, or very low-rate codes are used.