Massive MIMO: Pilot Design, Power Allocation, and Distributed Deployment

Abstract

Massive multiple-input multiple-output (MIMO) is widely acknowledged as the key enabling technology for the next-generation mobile communication networks. Pilot contamination (PC) is one of the key performance limiting factors for realizing the full potential offered by massive MIMO. This thesis investigates various aspects of PC in massive MIMO. First, we present methods to mitigate PC through pilot sequence design and location-aware pilot allocation. Second, we examine the impact of PC attack on the performance of massive MIMO. Third, we investigate practical aspects of massive MIMO deployment in wireless networks through distributed antenna arrays (DAAs).

In the first part of the thesis, we present two methods to mitigate PC. The first method, i.e., the pilot sequence design method, generates pilot sequences and devises power allocation at base stations (BSs) for downlink transmission. The pilot sequences and the proposed power allocation ensure that the predefined signal-to-interference-plus-noise ratio (SINR) requirements of all users are met. We derive new closed-form expressions for the user capacity and the user capacity region. Built upon these expressions, we develop an algorithm to obtain the required pilot sequences and power allocation.

The second method, i.e., the location-aware pilot allocation method, exploits the behavior of line-of-sight (LOS) interference among the users and allocates the same pilot sequence to the users with small LOS interference. Numerical results demonstrate that the proposed method significantly outperforms the existing methods. In the second part of the thesis, we present an active attack strategy in massive MIMO, where correlated pilots are used. The use of correlated pilots in massive MIMO has some intrinsic limitations, which can be exploited by an attacker to increase PC. To this end, we analyze the user capacity-achieving pilot sequence design for vulnerabilities that can be exploited by an active attacker. Furthermore, we present an effective active attack strategy. The active attacker transmits a known pilot sequence during the uplink training and artificial noise during the downlink data transmission phase. Numerical results reveal that the user capacity region of the network is significantly reduced in the presence of an active attack. Consequently, the SINR requirements of all the users are no longer guaranteed in the presence of active attackers.

In the third part of the thesis, we investigate downlink power control in DAA massive MIMO. We assume that the BS in each cell consists of multiple DAAs, which are deployed in arbitrary locations. Due to the spatial separation between antenna arrays, power control in DAA massive MIMO is a challenging problem as compared to conventional co-located massive MIMO. Based on the channel estimates obtained via uplink training, the BSs perform maximum ratio transmission for downlink. We then derive a closed-form spectral efficiency expression, where the channels are subject to correlated fading. Utilizing the derived expression, we propose a max-min power control algorithm to ensure that each user in the network receives a uniform quality of service.

The methods, algorithms, and techniques presented in this thesis give practical insights for solving challenges related to the deployment of massive MIMO in the next generation wireless networks. Show more