Loading...

Channel Estimation in Massive MIMO Systems in Modern Cellular Networks

Fozi, Mahdi | 2015

421 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 47618 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Hossein Khalaj, Babak
  7. Abstract:
  8. Telecommunications industry along energy industry will be among striking elements in improving next-generation lifestyle. With the growth of wireless communications at the heart of telecommunications industry, cellular networks as an inevitable part of our daily life, have been around as a fertile research field for both academia and industry. With the upcoming Heterogeneous Networks (HetNets), these researches are revisited to new horizons. After these shift paradigms,the proposed solutions, mostly fall into the category of network densification in terms of antenna per (active) user. Based on the strategy chosen to densify the network, we face two broad approaches: 1) Centralized densification: i.e., to equip the current macro base stations (BSs) with a very large number of antennas (100 or more) that can simultaneously accommodate many co-channel users, an idea referred to as massive multiple-input multiple-output (MIMO). 2) Distributed densification: i.e., small cells. Hence, on one hand, the well-established advantages of MIMO systems makes it the must element of the future telecommunications industry such that e.g., implementing up to 12 antennas is allowed in the next generation of Long Term Evolution-Advanced (LTE-A); enjoying from an uncorrelated-interference-free and noise-free regime, massive MIMO stands as the successor of the convenient MIMO systems. On the other hand, small cells, motivated by bringing the network closer to the users, have long attracted not only academia but also, even more, the industry. However, knowledge of the channel state information (CSI) at the BS side, is the key requirement for achieving the benefits of such systems.Traditional pilot-based approach for channel estimation either eats into bit rate of each user or results in pilot contamination, a phenomenon that severely degrades the performance of the system both in terms of the achievable bit rate and the experienced bit error rate. Motivated by the asymptotic orthogonality of the user’s channels in the massive regime, using the eigen vector decomposition of the autocorrelation matrix of the received signal, some blind channel estimation schemes have been proposed. Yet their performance is limited either by noise or interference in short channel coherence block. In this thesis, we propose a refined yet novel method to overcome this shortcoming. The proposed method utilizes the differential structure of the transmitted signal. However still the edge-cell users do not experience a reliable communication. Along with HetNets, it is suggested that small cells provide service to these set of users, to accompany the macro BSs. Finally simulation results solidify the derived theoretic results and outperformance of the proposed method relative to the current methods in the short coherence block setting
  9. Keywords:
  10. Channel Estimation ; Cellular Network ; Heterogeneous Networks ; Massive Multiple-Input Multiple-Output Systems ; Small Cells ; Channel Coherence Block ; Pilot Contaminations ; Asymptotic Orthogonality ; Differential Structure

 Digital Object List

 Bookmark

No TOC