A novel sparse multipath channel estimation model in OFDM system using improved Krill Herd-deep neural network.

Autor: Kondepogu, Vijayakumar, Bhattacharyya, Budhaditya
Zdroj: Journal of Ambient Intelligence & Humanized Computing; Mar2023, Vol. 14 Issue 3, p2567-2583, 17p
Abstrakt: Orthogonal Frequency Division Multiplexing (OFDM) is broadly used in modern wireless communication systems because of its highly robust nature when selecting the frequency of the wireless channels. In the case of coherent detection, channel estimation plays a significant role in the receiver design. The channel estimation is also required for the interference suppression or diversity combining, which has more receiver antennas. Channel estimation is pretended to be the heart of the OFDM-based wireless communication system. The pilot-aided channel estimation techniques based on frequency domain are performed according to the "Minimum Mean Square Error (MMSE) or Least Squares (LS)". LS-based methods are known to be less computationally complex and are also independent of prior Knowledge of Channel Statistics (KCS). But, the performance of the channel estimator in terms of Mean Square Error (MSE) encloses MMSE-based techniques that are superior to the incorporation of LS-based techniques. To improve the MSE performance, this paper presents the novel deep learning-based sparse multipath channel estimation in OFDM. Here, the enhanced Deep Neural Network (DNN) is used for enhancing the data communication performance via the estimated channel. The enhancement in DNN is done by the Adaptive Foraging Speed-based Krill Herd Algorithm (AFS-KHA), which helps to tune the parameters of DNN. This deep learning architecture demodulates the exact data through an estimated channel with less error that could overwhelm the "Matching Pursuit (MP) and Orthogonal MP (OMP) algorithms" that are general Compressed Sensing (CS) techniques. The analysis results reveal the better efficiency of the developed framework regarding the MSE and Bit Error Rate (BER) and also ensure less complexity in computation performance by comparing with the existing approaches. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index
Externí odkaz: