Reinforcement Learning for Link Adaptation in MIMO-OFDM Wireless Systems

Abstract

Abstract-Machine learning algorithms have recently attracted much interest for effective link adaptation due to their flexibility and ability to capture more environmental effects implicitly than classical adaptation algorithms. However, past applications are limited to rather simple configurations such as identifying channel condition or link adaptation in fixed or slowly varying channels. Recently, more sophisticated approaches using offline supervised learning have been proposed for link adaptation in complex configurations such as MIMO-OFDM. However, their time complexity and offline training phase hamper their real-world applicability. Approaches using online learning have shown good throughput performance, but the high memory requirement makes them inefficient or even impractical. In this paper, we propose a new effective online learning algorithm for link adaptation. Our computations show that the algorithm performs comparably to the existing online learning approaches, but ours requires minimal storage and time, which makes it more practical. Moreover it adapts to the change of channel distribution quickly.

Link adaptation for BICM-OFDM through adaptive kernel regression

“…None of these fulfills the constraints in Sec. III, but note that the kernel regression approach in [20] is actually quite close to the proposed QKLMS method in Sec. IV-B.…”

“…There have been various other proposals for link adaptation based on multidimensional features using kernel regression [20], [32], support vector machines [18], [19], [32], neural networks [21], or decision trees [11]. None of these fulfills the constraints in Sec.…”

“…Neither explicit loss differentiation nor explicit context awareness are necessary with this type of algorithm because it can adapt to changing environments. Another trend is to use multidimensional features beyond average SNR in order to obtain better predictors for the occurrence of packet errors [17], [18], [19], [20], [21]. Link adaptation strategies based on multidimensional features can be made context aware by suitable extension of the feature vector [11], [12], but most current strategies do not learn online.…”

Enhanced Adaptive Modulation and Coding (AMC) Technique Using ILDPCA Coders for MIMO Systems

“…Various coding schemes,Binary Phase Shift Keying (BPSK), Quadrature PSK (QPSK), 8-PSK, 8-QAM (Quadrature Amplitude Modulation), 16-PSK, 16-QAM, 64-PSK, 64-QAM etc. are used [5].However the exact number of environmental parameters like signal energy, noise variance, channel state information and other subtle factors like quantization error, made it hard to tune the transmission parameters optimally (even near optimal) [4]. Erik Johannesson et al [6] proposed the joint design of an optimal linear MIMO encoderdecoder pair of parallel Gaussian channels that can be formulated as a convex optimization problem followed by a matrix spectral factorization.…”

“…When Adaptive Modulation and Coding (AMC) scheme is used along with MIMO technology, it increases spectral efficiency by amending transmission parameters to the channel condition but still satisfying the target error performance [2]. The link adaptation techniques like adaptive modulation coding could dramatically improve the spectral utilization efficiency and throughput of wireless networks [3] [4]. Link-adaptation means adaptation of the code rate, the modulation constellation, and optionally, the powerallocation.…”