Complex Support Vector Machine Regression for Robust Channel Estimation in Lte Downlink System

Abstract

In this paper, the problem of channel estimation for LTE Downlink system in the environment of high mobility presenting non-Gaussian impulse noise interfering with reference signals is faced. The estimation of the frequency selective time varying multipath fading channel is performed by using a channel estimator based on a nonlinear complex Support Vector Machine Regression (SVR) which is applied to Long Term Evolution (LTE) downlink. The estimation algorithm makes use of the pilot signals to estimate the total frequency response of the highly selective fading multipath channel. Thus, the algorithm maps trained data into a high dimensional feature space and uses the structural risk minimization principle to carry out the regression estimation for the frequency response function of the fading channel. The obtained results show the effectiveness of the proposed method which has better performance than the conventional Least Squares (LS) and Decision Feedback methods to track the variations of the fading multipath channel.

Federated learning and next generation wireless communications: A survey on bidirectional relationship

“…77 In the earlier generation of wireless communication systems, channels are estimated through machine learning based shallow algorithms by formulating a regression problem received pilot signals as the input data and the channel state information (CSI) as the output data. In addition to this, support vector machine (SVM) algorithms have been used widely in literature for channel estimation [78][79][80][81][82] due to their ability to handle non-linear relationships between input and output data. Due to the increased complexity and nonlinearity in current 5G and future 6G wireless communication channels, deep learning evolved as a potential solution for channel estimation.…”

Artificial intelligence for channel estimation in multicarrier systems for B5G/6G communications: a survey

“…The main concept is to find a function that best fits the training data behavior to perform predictions. For example, linear [175][176][177], polynomial [178], 2D nonlinear [99,179,180], and support vector [172,173,[181][182][183][184][185][186][187] regressions have been employed in channel estimation for multicarrier systems. Regression algorithms go under the supervised learning paradigm.…”

“…The proposal was to map the input data into a finite-dimensional space to enable a higher-dimensional Hilbert space, similar to the approaches in [184,185]. A nonlinear SVR-based algorithm implemented with a radial basis function kernel for LTE systems leveraged the information in the pilot subcarriers to estimate the CFR [183]. The algorithm leads to lower BER under the same SNR compared with the LS and feedback estimators, from a good approximation to a perfect estimation.…”

Sparse representation in Szegő kernels through reproducing kernel Hilbert space theory with applications

“…We can use other loss functions and apply to them the corresponding analysis. In this section, we only consider the case using ε insensitivity cost function (5).…”

“…A comprehensive description of this method for classification and regression problems can be found in [4] and [34] respectively. It has been shown that SVMs perform well in practice in system identification, such as time series analysis with real SVM regression [23,35] and channel estimation in LTE Downlink system with complex SVM regression based on the Gaussian kernel [5]. In this paper, we will construct a complex SVM with the Szegő kernel which is complex-valued and never be used in SVM framework before.…”