We investigate a cell-free (CF) massive multi-input-multi-output (mMIMO) system wherein user equipments (UEs) employ generalized superimposed pilot (GSP) in the uplink, and are under a jamming attack. In GSP, pilots are superimposed onto data signals, which are precoded using a matrix orthogonal to that of pilots. This reduces pilot-signal interference in GSP transmission which, consequently, outperforms conventional superimposed pilot, and regular pilot transmissions in the jamming scenario. We also derive a closed-form spectral efficiency (SE) expression for this system, and use that to design a jamming-robust bilinear equalizer (JRBE) by using only the channel statistics. We show that JRBE design reduces the jammer impact on the attacked UE. We develop a percentile-based UE power optimization, which enhances the SE of a predefined percentile of UEs. This optimization, whose objective is a non-smooth function, is solved by developing a minorization maximization framework. We show that in a jammed system, GSP-JRBE combination can provide a high SE to large number of UEs over existing designs. We also show that percentile optimization can enhance the SE of the jammed UE, without sacrificing the SE of unjammed UEs.

We consider the uplink of a hardware-impaired intelligent reflective surfaces (IRS) aided multi-cell massive multiple-input multiple-output (mMIMO) system with mobile user equipments, whose channel age with time. For this system, we design a novel distortion-and-aging-aware MMSE (DAAMMSE) receiver that not only provides a higher spectral efficiency (SE) than conventional maximal ratio and distortionunaware MMSE (DU-MMSE) receivers, but also reduces the pilot overhead. We develop a novel low-complexity IRS phase optimization framework based on minorization-maximization (MM) technique, which requires only channel statistics to calculate the optimal phase. We also show that the SE gain of the DAA-MMSE receiver over DU-MMSE receiver increases with hardware impairments, and channel aging. Along with DAAMMSE receiver, the IRS is also shown to reduce the pilot overhead in a mMIMO system with channel aging.

We consider an unmanned aerial vehicle (UAV)-relay-assisted massive multi-input multi-output system, where a base station (BS) serves coverage-limited user equipments (UEs) using non-orthogonal multiple access (NOMA). Due to the UAV altitude, the channels contain both line-of-sight (LoS) and non-LoS components. Moreover, the mobility of UAVs and UEs causes channel aging. We analyse the system spectral efficiency (SE) with minimum mean square error precoding, and by considering practical Rician channels with phase-shifts, channel aging, imperfect successive interference cancellation, and pilot contamination. We also develop an iterative stochastic optimization algorithm to optimize the system global energy efficiency. Our algorithm has extremely low complexity, due to its closed-form power updates, and converges faster than the existing state-of-the-art deterministic algorithm, which numerically calculates the SE.

We consider the uplink of a hardware-impaired intelligent-reflective surfaces (IRS)-aided multi-cell massive multiple-input multiple-output (mMIMO) system with mobile user equipments, whose channels age with time. We propose a novel distortion-and-aging-aware minimum-mean-square-error (DAA-MMSE) receiver that not only provides a higher spectral efficiency (SE) than conventional maximal ratio combiner (MRC) and distortion-unaware MMSE (DU-MMSE) receivers, but also reduces the pilot overhead. We also derive an SE lower bound for this system with MRC and DAA-MMSE receivers. We develop a novel low-complexity SE optimization framework using the weighted-MMSE and minorization-maximization (MM) techniques, which provide closed-form expressions for optimal power and phase allocations. The MM technique requires construction of a surrogate function, which we do in this work. We numerically characterize the i) UE velocities for which IRS deployment can tangibly increase their SE; ii) reduced pilot overhead with DAA-MMSE receiver along with IRS.

This work considers a device-to-device (D2D) underlaid multi-cell massive multi-input multi-output (mMIMO) system, wherein multiple cellular users (CUs) and D2D user-pairs support ultra-reliable low-latency communications (URLLC). The CUs are served by the mMIMO base station (BS), while the D2D user-pairs communicate without BS intervention. We design a novel interference-aware maximal ratio (IA-MR) combiner, and show that it has a similar spectral efficiency (SE) as the zero forcing (ZF) combiner, and that too with a much lesser computational complexity. The IA-MR combiner recovers the SE loss due to URLLC and D2D underlaying in multi-cell mMIMO systems. We also derive closed-form SE expression for the IA-MR combiner by assuming spatially-correlated Rician-faded channels, and use them to maximize the global energy efficiency (GEE) metric. The URLLC and D2D operations cause the GEE to become non-convex, as it now contains difference-of-concave and coupled fractional functions of optimization variables. We propose an optimization framework to tackle this non-convexity by using the quadratic and Lagrangian dual transforms. The proposed framework is also clubbed with a stochastic technique to optimize the GEE for ZF and minimum mean squared error equalizers, which do not have closed-form SE expressions.

We consider the uplink of a cell-free (CF) massive multi-input multi-output (MIMO) system with superimposed pilot (SP) transmission, wherein user equipments (UEs) superimpose low-powered pilots onto data signals. This is unlike regular pilot (RP) transmission, where data and pilots use orthogonal spectral resources. Our CF mMIMO system has hardware impairments which occur due to i) low-quality radio frequency (RF) chains at the access points (APs) and UEs; and ii) dynamic analog-to-digital converter (ADC) architecture at the APs, which enables each RF chain to be connected to different resolution ADC. We derive a closed-form spectral efficiency (SE) expression for this CF system, wherein UEs observepracticalspatially-correlated Rician-faded channels. The derived lower-bound is generic, and reduces to the ones in the existing CF mMIMO SP works,which have only considered ideal hardware. Using this lower-bound, we optimally balance pilot and data transmit powers to maximize the SE. We analytically show that the optimal power balance is insensitive to AP impairments, but sensitive to that of UEs. We numerically show that SP can provide a higher SE than RP for low-to-severe hardware impairment levels, when supported by dynamic ADC architecture at the APs. With low resolution ADCs, RP always outperforms SP. The RP is also shown to be suitable for low UE speeds.

We consider two-way full-duplex (FD) massive multiple-input multiple-output decode-and-forward (DF) relaying with ultra-reliable low-latency communication (URLLC) users. The FD relay employs hybrid architecture with a fewer radio frequency (RF) chains than the antennas. Further, the users observe spatially-correlated Rician channels. For this system, we propose a novel optimal bilinear equalizer (OBE) that not only provides a much higher spectral efficiency (SE) than the conventional maximal ratio (MR) combiner, but also has better multi-user-interference (MUI) cancellation capability, which translates to low latency, as the system can schedule a large number of URLLC users. The OBE achieves this gain by exploiting spatial correlation in the system, an aspect we analytically demonstrate. We also derive a SE lower bound for this system, which is used to maximize the global energy efficiency (GEE) metric. The objective of GEE is a fractional function of the optimization variables, and is non-convex. We extend the fractional programming tools – Lagrangian dual and Quadratic transforms – to optimize the GEE. We show that a hybrid DF relay has a higher GEE, and a similar SE as that of a digital amplify-and-forward relay. We also show the higher GEE provided by a combination of OBE design and hybrid architecture, when compared with conventional combiners and digital architecture.

We consider the downlink of a cache-aided cell-free (CF) massive multi-input multi-output (mMIMO) system, wherein user equipments (UEs) exploit cached data to cancel multi-user interference (MUI) in the downlink. This work shows the importance of downlink pilots transmitted by the APs, which are used by the UEs to first estimate the instantaneous downlink channel state information (CSI), and then perform cache-aided MUI cancellation. We derive a closed-form SE expression for this system by considering practical spatially-correlated Rician fading channels with phase shifts, pilot contamination, and hardware impairments, both at the APs and UEs. We also propose a joint cache placement and power control optimization to maximize the global energy efficiency (GEE) metric. The cache placement problem is solved via clustering, while the power is optimized by designing a novel low-complexity parallel block minorization-maximization optimization. We numerically validate the benefits of downlink CSI for cache-aided CF mMIMO systems, and also show that degradation in CSI quality dramatically negates the UEs cache-aided MUI cancellation capability. We also show that the proposed optimization yields the same GEE as existing state-of-the-art optimizations, but with a much lower complexity.

We consider a multi-way massive multi-input multi-output (mMIMO) relaying system wherein a relay aids multi-way data exchange between multiple users by employing non-orthogonal multiple access (NOMA). The relay achieves this by superposing different user signals in the power domain. Each user then sequentially decodes the data of all other users by performing successive interference cancellation (SIC). To perform SIC, a user utilizes the downlink channel information, which it estimates using the precoded pilots transmitted by the relay. We derive a closed-form spectral efficiency (SE) expression for this downlink-pilot-aided NOMA multi-way relaying system by considering spatially-correlated channels, and imperfect SIC at the user. We next design two algorithms to maximize this SE by optimally allocating the user transmit powers and the NOMA variables. Both these algorithms provide a similar SE, but the latter one, due to its closed-form power updates, has a much lesser complexity. The efficacy of downlink pilots in multi-way NOMA relaying, with practical correlated Rayleigh-faded channels, is shown by demonstrating that they vastly outperform the case when users perform SIC using channel statistics.

We consider the downlink of a cell-free (CF) massive multi-input multi-output (mMIMO) system, wherein multiple access points serve clustered users by employing non-orthogonal multiple access (NOMA) technology. This work shows the importance of transmitting downlink pilots for a NOMA CF mMIMO system, and derives a closed-form spectral efficiency (SE) expression with spatially-correlated Rician channels by addressing the complexity due to i) dynamic-resolution analog to digital converters; and ii) channel estimation errors. We also design a low-complexity parallel block minorization-maximization (MM) algorithm to optimize the non-convex global energy efficiency (GEE) metric. The MM algorithm requires a surrogate function for the GEE metric, which we construct in this work. We show that the NOMA CF mMIMO system, wherein channels estimated using downlink pilots are used for performing successive interference cancellation (SIC), significantly outperforms its counterpart which uses channel statistics to perform SIC. We also show that our parallel block MM optimization yields the same GEE as an existing optimization, but with a much lower complexity.

We study the impact of channel aging on the uplink of a cell-free massive multiple-input multiple-output system with hardware impairments. We consider a dynamic analog-to-digital converter architecture at the access points (APs), and low-resolution digital-to-analog converters at the user equipments (UEs). We derive a closed-form spectral efficiency expression by considering i) practical spatially-correlated Rician channels; ii) hardware impairments at the APs and the UEs; iii) channel aging; and iv) large-scale fading decoding (LSFD). We show that LSFD can effectively mitigate the detrimental effects of i) channel aging for both low and high UE velocities; and ii) inter-user interference for low-velocity UEs but not for high-velocity UEs.

We consider a hardware-impaired multi-cell Rician-faded massive multi-input multi-output (mMIMO) system with two-layer pilot decontamination precoding, also known as large-scale fading precoding (LSFP). We derive a closed-form spectral efficiency (SE) expression by assuming a flexible dynamic analog-to-digital converter (ADC)/digital-to-analog converter (DAC) architecture, and hardware-impaired radio frequency chains at the base stations (BSs) and user equipments. The dynamic ADC/DAC architecture enables us to vary the resolution of ADC/DAC connected to each BS antenna, and suitably choose them to maximize SE. We design a distortion-aware minimum mean squared error (DA-MMSE) precoder, and investigate its usage by combining it with the two-layer LSFP, and conventional single-layer precoding (SLP). We show that the DA-MMSE precoder with SLP outperforms its distortion-unaware counterpart, which is used with LSFP. We analytically show that for pure LoS channels, the LSFP reduces to SLP, and its implementation can thus be avoided. It is shown that the LSFP can tolerate high hardware impairments at the BS, but is extremely sensitive to the ADC resolution of the user. We also optimize the global energy efficiency by using a minorization-maximization based algorithm, and show its improved performance over the conventional SE optimization techniques.

We consider a full duplex (FD) multi-cell massive multi-input-multi-output (mMIMO) network, wherein FD mMIMO base-stations (BSs) serve multiple FD user equipments (UEs) on the same time-frequency resource. We derive closed-form uplink and downlink spectral efficiency (SE) expressions by considering i) radio-frequency impairments at the BSs and the UEs; ii) dynamic-resolution analog-to-digital converter/digital-to-analog converter (ADC/DAC) architecture at the BSs and low-resolution ADC/DAC at the UEs; and iii) spatially-correlated Rician channels. We optimize the non-convex global energy efficiency metric for this FD system by using the minorization-maximization (MM) framework, which transforms the optimization into a sequence of surrogate concave problems. We construct a novel surrogate function to exploit the MM framework. We numerically characterize the effect of FD interference on Rayleigh- and Rician-faded mMIMO networks, and show the higher immunity of the latter to the FD interference. We also show the SE gain provided by the dynamic ADC/DAC architecture which allows us to judiciously choose ADC/DAC resolution, while reducing the power consumption.

We study the impact of channel aging on the uplink of a cell-free (CF) massive multiple-input multiple-output (mMIMO) system by considering i) spatially-correlated Rician-faded channels; ii) hardware impairments at the access points and user equipments (UEs); and iii) two-layer large-scale fading decoding (LSFD). We first derive a closed-form spectral efficiency (SE) expression for this system, and later propose two novel optimization techniques to optimize the non-convex SE metric by exploiting the minorization-maximization (MM) method. The first one requires a numerical optimization solver, and has a high computation complexity. The second one with closed-form transmit power updates, has a trivial computation complexity. We numerically show that i) the two-layer LSFD scheme effectively mitigates the interference due to channel aging for both low- and high-velocity UEs; and ii) increasing the number of AP antennas does not mitigate the SE deterioration due to channel aging. We numerically characterize the optimal pilot length required to maximize the SE for various UE speeds. We also numerically show that the proposed closed-form MM optimization yields the same SE as that of the first technique, which requires numerical solver, and that too with a much reduced time-complexity.

This work considers a multi-way massive multi-input multi-output (mMIMO) relaying system wherein several wireless information and power transfer capable users, first harvest energy, and then exchange information via a relay. The mMIMO relaying system is practically modelled by considering spatially-correlated relay-user channels, and by using a non-linear energy harvesting (EH) model. A closed-form spectral efficiency expression is derived for this system, which is then used to jointly allocate the relay and users transmit powers, and the charging time to optimize the weighted sum energy efficiency (WSEE) metric. The spatially-correlated channel and the non-linear EH model not only couple the optimization variables, but also introduce nested scalar fractional optimization functions. This makes the WSEE metric non-convex. The proposed solution first uses the existing quadratic transformation (QT) to decouple the scalar fractional forms. It then applies novel results, which are developed by extending the QT framework, to decouple the optimization variables. The proposed WSEE optimization algorithm is shown to provide close-to-optimal WSEE. It is also shown that channel spatial correlation increases the amount of energy harvested by the users, but reduces their WSEE.

We consider a practical hardware-impaired multi-cell massive multi-input multi-output (mMIMO) system with spatially-correlated Rician fading channels, whose line-of-sight component experiences a random phase-shift. The multi -cell mMIMO system employs two-layer large-scale fading decoding (LSFD) to mitigate the interference due to pilot contamination. We consider a dynamic analog-to-digital converter (ADC) architecture at the base stations (BSs), which enables us to independently vary the resolution of ADC connected to each of its antenna, and thus properly choose them to achieve a high spectral efficiency (SE). Both BS and user equipments (UEs) are also equipped with low-cost radio frequency chains, which introduce additional hardware impairments. We first derive a closed-form SE expression for this system, and then optimize the LSFD vectors to maximize the sum SE metric using generalized Rayleigh quotient. We investigate the ADC resolution and hardware impairments values for which LSFD is highly effective, and the values for which its effectiveness reduces.

We consider a hardware-impaired multi-cell Rician- faded massive multi-input multi-output (mMIMO) system with two-layer pilot decontamination precoding, also known as large-scale fading precoding (LSFP). Each BS is equipped with a flexible dynamic analog-to-digital converter (ADC)/digital-to-analog converter (DAC) architecture and the user equipments (UEs) have low-resolution ADCs. Further, both BS and UEs have hardware-impaired radio frequency chains. The dynamic ADC/DAC architecture allows us to vary the resolution of ADC/DAC connected to each BS antenna, and suitably choose them to maximize the SE. We propose a distortion-aware minimum mean squared error (DA-MMSE) precoder and investigate its usage with two-layer LSFP and conventional single-layer precoding (SLP) for hardware-impaired mMIMO systems. We discuss the use cases of LSFP and SLP with DA-MMSE and distortion-unaware MMSE (DU-MMSE) precoders, which will provide critical insights to the system designer regarding their usage in practical systems.

We consider two-way massive multiple-input multiple-output (mMIMO) relaying with ultra-reliable low-latency communication (URLLC) users. The full-duplex (FD) relay employs hybrid architecture, with a lesser number of radio frequency chains than the antennas. For this system, we first propose a novel digital optimal bilinear equalizer which not only provides a much higher spectral efficiency (SE) than the conventional maximal ratio combiner, but also yields a lower latency when a large number of URLLC users are scheduled. We also derive a SE lower bound for this system by considering spatially-correlated Rician fading channels. We also show that the hybrid decode-and-forward relay provides much higher SE to its URLLC users than a full-RF chain amplify-and-forward relay, which only serves non-URLLC users.

We consider a relay-aided massive multi-input multi-output (mMIMO) system where a base station (BS) serves various users via multiple relays by employing non-orthogonal multiple access (NOMA). We practically model this system by considering spatially-correlated Rician-faded channels, and channel estimation errors both at the BS and at the users, which consequently perform imperfect successive interference cancellation (SIC). We consider these two artifacts, and derive a lower bound on the spectral efficiency (SE) of this multi-relay NOMA system, which is valid for arbitrary number of BS antennas. We crucially show that i) Rician-faded channels are immune to the errors caused by imperfect SIC at users; ii) high spatial correlation improves the spatial diversity among relays located close to BS and achieves better sum SE than its uncorrelated counterpart.

We consider the downlink of a relay-aided multi-cell massive multi-input multi-output (mMIMO) system, where in each cell the base station (BS) serves its users via multiple relays by employing non-orthogonal multiple access (NOMA). We model this system by considering spatially-correlated Rician-faded channels and their estimation errors. The users, consequently, perform imperfect successive interference cancellation (SIC). We derive a lower bound on the spectral efficiency (SE) of this system. We then optimize the non-convex global energy efficiency (GEE) metric, which is a fractional function of the optimization variables. We solve this problem by considering a low-complexity alternating minimization maximization approach, which splits a complex joint problem into multiple simpler convex surrogate sub-problems. We propose a novel surrogate function to exploit this framework, and analytically show that it satisfies the desirable properties of a valid surrogate function. We numerically show that 1) reusing the pilots in each cell, when the channel has sufficiently hardened, provides higher SE than using orthogonal pilots in all cells and 2) the proposed GEE algorithm provides similar GEE as that of an existing joint optimization framework, but with much less complexity.

We consider a two-way full-duplex (FD) system where a massive multi-input-multi-output FD base station (BS) communicates with multiple FD users via an intelligent reflecting surface (IRS). We derive a closed-form network spectral efficiency lower bound by considering spatial correlation at the BS and IRS. We demonstrate the importance of modelling spatial correlation by simplifying this lower bound for uncorrelated channels to show that the IRS capability to modify the wireless medium now is significantly impeded. This lower bound, which is a function of only the long-term channel statistics, is also used to maximize the non-concave global energy efficiency metric by optimally allocating the transmit powers, and by designing the IRS phases. We derive closed-form updates for optimizing the transmit powers using Lagrangian dual, Quadratic, and Dinkelbach’s transforms. We then optimize the IRS phases by using the projected gradient ascent algorithm. We numerically show that increasing the number of IRS elements can help a FD mMIMO BS outperform its half-duplex counterpart, which otherwise under-performs due to its limited ability to cancel various FD interferences.

We consider a cell-free (CF) massive multi-input multi-output (mMIMO) system, where multi-antenna access points (APs) serve single-antenna unmanned aerial vehicles (UAVs) and ground users (GUEs). We assume, unlike the existing CF mMIMO literature, hardware-impaired UAVs and GUEs, which observe a mixture of spatially-correlated Rician- and Rayleigh-faded channels while communicating with hardware-impaired APs. We derive a closed-form downlink spectral efficiency (SE) expression by using practical models for the channel mixture, and by considering channel estimation errors. We propose a novel block quadratic transformation (block-QT) technique to optimize non-convex network-centric global energy efficiency (GEE) by appropriately modeling circuit, UAV propulsion and fronthaul powers. The novel block-QT approach combines block optimization and quadratic transformation technique to decompose GEE optimization into simpler convex sub-problems. We numerically show that i) it is better to operate a UAV at a larger height when it has severe hardware impairments; and ii) when UAVs operate at a lower height, they do not significantly affect the SE of GUEs.

We consider a multi-user system where a massive multi-input-multi-output (mMIMO) full-duplex (FD) base-station (BS) communicates with multiple FD users via an intelligent reflecting surface (IRS). We derive novel uplink and downlink spectral efficiency (SE) lower bound expressions when the BS estimates composite BS-IRS-user channels. The lower bounds are derived considering spatially-correlated IRS and user channels, and require only statistical channel state information (CSI). We propose a projected gradient ascent based IRS phase optimization algorithm, which also uses only statistical CSI, and enables the system to achieve a higher SE in the presence of the loop and inter-user interferences. We analytically investigate the dependence of SE on the IRS phase for spatially correlated channels considered herein. We show that the proposed analysis can help in increasing the SE by appropriate IRS placement.

We consider a full-duplex cell-free massive multiple-input-multiple-output system with limited capacity fronthaul links. We derive its downlink/uplink closed-form spectral efficiency (SE) lower bounds with maximum-ratio transmission/maximum-ratio combining and optimal uniform quantization. To reduce carbon footprint, this paper maximizes the non-convex weighted sum energy efficiency (WSEE) via downlink and uplink power control, and successive convex approximation framework. We show that with low fronthaul capacity, the system requires a higher number of fronthaul quantization bits to achieve high SE and WSEE. For high fronthaul capacity, higher number of bits, however, achieves high SE but a reduced WSEE.

We consider a multi-way massive multi-input multi-output (mMIMO) relay via which several energy-constrained MIMO users exchange information by switching between wireless power generation and information transfer. The mMIMO relay and MIMO users experience spatially-correlated channels, which the current multi-way relaying literature ignores. We derive closed-form spectral efficiency (SE) expression for this system that is valid for a practical number of relay antennas. We design a max-min power control algorithm which ensures user fairness by jointly optimizing the charging time and transmit powers of the relay and users.

We consider an unmanned aerial vehicle (UAV) enabled cell-free (CF) massive multi-input multi-output (mMIMO) system, where multi-antenna access points (APs) assist various single-antenna UAVs and ground users (GUEs). Unlike, existing CF mMIMO works, we assume non-ideal transceiver hardware at the UAVs, GUEs and APs, and derive a closed-form downlink spectral efficiency (SE) expression by considering spatially-correlated Rician channels and channel estimation errors. We also propose a novel block quadratic transformation technique to optimize the system global energy-efficiency (GEE) by incorporating practical backhaul power, circuit and UAV propulsion power. We numerically show that severely hardware-impaired UAVs will have a higher SE when they operate at a larger height.

Most of the existing massive multi-input-multi-output (mMIMO) relaying works assume ideal transceiver hardware and spatially uncorrelated channels. A practical mMIMO relay, however, due to limited antenna spacing and the use of cost-effective equipment, usually experiences spatially-correlated channels and hardware impairments, respectively. We consider a hardware-impaired two-way full-duplex (FD) spatially-correlated mMIMO relaying with multiple MIMO FD user-pairs, and derive a closed-form spectral efficiency (SE) expression which is valid for a practical number of finite relay antennas. We use this expression to develop a quadratic transformation approach to optimize the non-convex global energy efficiency (GEE) and weighted sum energy efficiency (WSEE) metrics, which are fractional functions of optimization variables. The proposed approach first converts the fractional problem into its equivalent fraction-free counterpart, and then uses novel transformations to make the problem concave, and solve it using an iterative algorithm. We use this optimization to i) investigate the impact of spatial correlation and hardware impairments on GEE; ii) decide hardware impairment values for a tolerable reduction in GEE; and iii) investigate the impact of weights in WSEE on the users energy efficiency.

We optimize the weighted sum energy efficiency (WSEE) of two-way amplify and forward relaying, where multiple full-duplex (FD) user-pairs exchange information via a shared FD massive multiple-input multiple-output (MIMO) relay. Optimization of user-centric WSEE metric, which prioritizes links of users with high energy efficiency (EE) requirements by suitably choosing their weights, is a non-convex problem due to its sum-of-ratio form. We optimize it by first approximating WSEE as a concave-convex fractional function, and then by using the quadratic transform. We then use Karush-Kuhn-Tucker (KKT) conditions to derive a closed-form solution to optimize WSEE. We numerically show that the i) proposed solutions achieve significant WSEE gains; and ii) suitable choice of weights can help prioritize EE requirements of different users.

We consider a multi-pair two-way full-duplex (FD) amplify-and-forward relaying, where multiple FD multi-input multi-output (MIMO) users exchange information via a shared FD massive MIMO relay. We derive a closed-form spectral efficiency (SE) lower-bound for maximal-ratio combining/maximal-ratio transmission relay processing considering spatially correlated relay and user antennas, which most of the existing works have ignored. The derived SE lower-bound is applicable for arbitrary number of relay antennas, and simplifies to the following results available in the literature i) the asymptotic SE for number of relay antennas tending to infinity; and ii) the SE lower-bound with independent relay antennas and single-antenna users. We use this SE lower-bound to maximize the non-convex SE, which is of matrix fractional form. We optimize SE by approximating it as concave-convex function, and then by applying matrix quadratic transformation. We numerically validate the SE lower-bound for various system configurations and also characterize their performance for different spatial correlation and interference values. We investigate the gains achieved by the optimal SE over equal power allocation.