Measurement and modeling of multiuser multiantenna system in aircraft in the presence of electromagnetic noise and interference

This paper evaluates the accuracy with which the performance of a multi-user multi-antenna system can be predicted with and without considering co-channel interference and noise (Gaussian, α- stable and Cauchy) using a site-specific 3D ray-tracing algorithm as well as with statistical models with Gaussian and Nakagami-m channel models in small to medium sized aircraft. These models expand on previous statistical channel models such as the hyper-Rayleigh model by including the simultaneous effects of co- and adjacent channel interference, antenna matching, efficiency, directivity and polarization as well as (for the 3D model) site-specific multipath effects. Measurements and comparisons are made in a metallic-bodied Beech Baron BE 58P and a composite structure Rockwell T-39 Sabreliner. It was found that the 3D ray tracing model provides a mean capacity within 1 % of those measured in the two aircraft in the presence of interference and noise. This was closely followed by the Nakagami-m distribution (m=1.4) which was within 1-3% of measured capacity in the presence of interference and within 6% for a combination of interference and noise and the Gaussian model which was within 6% of measured capacity in the presence of interference and within 11% for a combination of interference and noise . The Cauchy noise degraded the capacity more than the other types of noise in the aircraft, providing a lower bound for capacity in an aircraft system.