Article Preview
Top1. Introduction
Millimeter (MM) wave communication is a growing technology in wireless systems. The desire for enhanced data rates to support high quality multimedia applications with lower latency levels can be realized using MM wave communications. Over 20 GHz of unused spectrum exists in the 28, 38 and 72 GHz bands which can be utilized for serving increasing number of users in cellular networks (Hur et al., 2013). This amount of spectrum is so huge that many more mobile users can be accommodated with higher data rates for all the modulation schemes. The large amount of unused spectrum in the millimeter-wave frequency bands can be utilized to yield hundred times more capacity than that of 4G cellular networks (Janssen et al., 1994). In order to achieve enhanced data rates and channel capacity, MM wave communications have to employ high gain directional antennas. Higher directional gain can be obtained by deploying phased antenna array systems. As already mentioned, the MM wave frequencies are in higher range of GHz, the antennas will thus be of sizes in the range of millimeter such that they form phased antenna arrays when integrated on chip. It is important to understand that the phased array output SNR at the receiver plays a crucial role in channel capacity and data rate analysis of cellular systems (Andrews et al., 2001). Recent studies have focused on effective MM wave channel modeling for capacity enhancement in such links (Janssen et al., 1994). Some closely related works are discussed here.
Heath et. al. focused on the challenges associated with signal processing in MM wave wireless systems. The channel model equations for both 2-D antenna and 3-D antenna arrays with beamsteering were proposed. The representation in terms of beam space was also presented for an N -element Uniform Linear Array (ULA) system (Heath, 2016). The adverse effect of inter antenna correlations on performance of MIMO systems is quite prominent at MM wave frequencies (Janssen et al., 1994) (Heath, 2016). Effective analytical modeling of MM wave channels is a key to control such correlations (Hur et al., 2013). Towards this, the work presented in (Mckay et al., 2007) utilizes Kronecker modeling to analyze conventional MIMO systems in terms of inter-antenna correlation separately at the transmitter and receiver end. Similar studies are proposed in (Kshetrimayum, 2017) which analytically express the dependance of crucial MIMO channel parameters on antenna correlations. Additionally, the variation in performance levels due to Equicorrelated modeling of such systems is stressed upon in (Kshetrimayum, 2017). Hamadeh et. al. analyzes Massive MIMO system performance under correlated MM wave environments using the popular Saleh Valenzuela (SV) physical channel model (Molisch, 2004) in (Hemadeh et al., 2018). Performance evaluation is carried out primarily under LOS scenarios. Similarly, Bjornson et. al. analyzes Correlated Massive MIMO systems with sub-6 GHz as well as MM wave carriers in (Bjornson et al., 2019). It is noteworthy to mention that introducing antenna correlations only improves the link performance marginally in both (Hemadeh et al., 2018) and (Bjornson et al., 2019). In most cases, MM wave propagations consist of LOS paths where it is typical to consider physical models as they reflect channel parameters accurately (Hemadeh et al., 2018). However, recently proposed works in (Khatun, 2017)-(Rappaort, 2017) stress upon MM wave channel characterization of NLOS links. Contrary to physical models, the precise impact of antenna correlations on such systems are stochastically analyzed utilizing Kronecker MIMO decompositions (Kshetrimayum, 2017) (Rappaort, 2017). Very few works discuss the possibility of tuning MM wave system performance by controlling correlation levels. In this direction, the current work initially describes how MIMO taps and correlation matrix entries only in the MM wave domain are dependent on steering vectors or corresponding departure and arrival angles (Heath, 2016). This dependency does not exist in case of microwave links (Mckay et al., 2007). As such, effective tuning of steering angles is utilized to obtain minimal equicorrelations at both transmitter and receiver ends. Based on it, this paper further proposes a novel Minimally Equicorrelated MIMO- MRC model which achieves capacity improvements in short range NLOS MM wave links. The key contributions of the paper are highlighted as below: