Error Exponents and Cutoff Rate for Noncoherent Rician Fading Channels

TL;DR

This paper investigates error exponents and cutoff rate for noncoherent Rician fading channels without channel state information, revealing optimal input distributions under various power constraints and analyzing low-power behavior.

Contribution

It derives bounds on error exponents and characterizes optimal input distributions for noncoherent Rician channels under different power constraints.

Findings

Optimal input has discrete amplitude and uniform phase.

Optimal input remains discrete under peak power constraints.

Single-mass input is optimal at low power for cutoff rate.

Abstract

In this paper, random coding error exponents and cutoff rate are studied for noncoherent Rician fading channels, where neither the receiver nor the transmitter has channel side information. First, it is assumed that the input is subject only to an average power constraint. In this case, a lower bound to the random coding error exponent is considered and the optimal input achieving this lower bound is shown to have a discrete amplitude and uniform phase. If the input is subject to both average and peak power constraints, it is proven that the optimal input achieving the random coding error exponent has again a discrete nature. Finally, the cutoff rate is analyzed, and the optimality of the single-mass input amplitude distribution in the low-power regime is discussed.