Highly accurate /spl eta/-μ approximation to the sum of M independent nonidentical Hoyt variates

Sums of Hoyt random variables occur in several wireless communication applications. Notwithstanding its practical interest, no closed-form expression exists for the sum statistics, which must be often evaluated via intricate multifold numerical integration. We use the /spl eta/-μ distribution to approximate the distribution of the sum of M independent, nonidentically distributed Hoyt random variables having arbitrary mean powers and arbitrary fading degrees. To this end, appropriate /spl eta/-μ distribution parameters are derived. Our results show that the differences between exact and approximate statistical curves are almost imperceptible.