Does Fast Adaptive Modulation Always Outperform Slow Adaptive Modulation?

Abstract

Abstract-Link adaptation techniques are important modern and future wireless communication systems to cope with quality of service fluctuations in fading channels. These techniques require the knowledge of the channel state obtained with a portion of resources devoted to channel estimation instead of data and updated every coherence time of the process to be tracked. In this paper, we analyze fast and slow adaptive modulation systems with diversity and non-ideal channel estimation under energy constraints. The framework enables to address the following questions: (i) What is the impact of non-ideal channel estimation on fast and slow adaptive modulation systems? (ii) How to define a proper figure of merit which considers both resources dedicated to data and those to channel estimation? (iii) Does fast adaptive always outperform slow adaptive techniques? Our analysis shows that, despite the lower complexity and feedback rate, slow adaptive modulation (SAM) can achieve higher spectral efficiency than fast adaptive modulation (FAM) in the presence of energy constraint, diversity, and non-ideal channel estimation. In addition, SAM satisfies bit error outage requirements also in FAM-denied region.

Adaptive OFDM for Wireless Interconnect in Confined Enclosures

“…The preamble consists of synchronization sequences followed by six pilot symbols for channel estimation essential for channel equalization. With the received signal strength intensity and received signal quality indexes also foreseen by the standard, the receiver can determine the instantaneous SNR [20] of individual subcarriers.…”

Fast Decoding and Hardware Design for Binary-Input Compressive Sensing

“…Toni and Conti [16] find that, under nonideal channel estimation, fast adaptive modulation (FAM) techniques do not perform as well as slow adaptive modulation (SAM) techniques. Duel-Hallen et al [17] point out that, in order to realize the potential of adaptive transmission methods, fading channel variations have to be reliably predicted at least several milliseconds in advance.…”

“…Since the LLR message of each bit is available in the previous iteration, the probability is defined as (15) where is the flip of taking binary value too. Let us further define the probability of equal to the vector in the th row of Table_l_k as (16) Substituting (15) into (16), the probability can be written as (17) Equation (17) can be calculated in the log domain (18) The first term is easy to calculate. Similar to the message passing algorithms for Turbo codes and LDPC codes, the second term in (18) can be calculated by looking up the table of or approximated with a piece-wise function [52].…”

Rate-adaptive modulation for square OSTBCs in arbitrarily correlated Rayleigh fading with imperfect channel estimation

“…The ASE is defined as η times the SE, where η K/(K + K p ) represents the fraction of the resources that is used for the transmission of the datadependent portion of the frame [13]. Hence, the average ASE (in bits/s/Hz) is obtained as…”