Turbo code decoding in MAI environment

Interference in several situations, like multiple access interference (MAI) in dense sensor networks, shows strong but rarely occurring pulses that greatly disturb communication, leading to serious performance degradation in classical decoders based on the Gaussian assumption. We propose to represent this impulsive noise with symmetric α-stable (SαS) distribution, well adapted due to its heavy tailed property. Therefore, the Euclidean distance derived from the Gaussian distribution is no longer valid for the SαS environment. We attempt to use the p-norm as a distance metric and evaluate it in the decoding scheme of turbo codes. Based on our simulation results, the new metric significantly reduces the bit error rate (BER) compared with the decoders using the Euclidean distance. It indicates that the p-norm could be an appropriate metric adapted to the MAI environment for turbo code decoding.