To enable improved spectral efficiency and uniform service of cell-free massive multiple-input multiple-output (CF-mMIMO) systems, challenges such as excessive communication overheads for gathering global CSI at the central processor (CP) and high computational complexity for centralized interference management should be addressed. Deep learning (DL)-based decentralized optimization approaches have been proposed to tackle these challenges. However, conventional black-box deep neural network (DNN)-based fronthaul coordination methods require large DNNs and are less adaptable to changes in wireless channel statistics and environments. In this paper, we develop a model-based neural collaboration (NC) framework, which leverages DNN-assisted signal processing built upon our proposed low-complexity distributed weighted minimum mean squared error (L-D-WMMSE) algorithm, for downlink CF-mMIMO beamforming. DNN components for neural initialization (NI) and neural parameterization (NP) reduce the fronthaul overheads while improving the sum-rate of user equipments (UEs). Through our model-based approach, these DNN components are designed to be compact, thereby reducing computational complexity. We derive gradients for DNN updates that each AP can compute using its local CSI and the fronthaul messages exchanged for downlink beamforming. This enables a distributed online DNN adaptation that runs during normal operation of the CF-mMIMO system at zero additional fronthaul overheads.

This work studies the uplink of a multi-tenant cloud radio access network (C-RAN) system with spectrum pooling. In the system, each operator has a cloud processor (CP) connected to a set of proprietary radio units (RUs) through finite-capacity fronthaul links. The uplink spectrum is divided into private and shared subbands, and all the user equipments (UEs) of the participating operators can simultaneously transmit signals on the shared subband. To mitigate inter-operator interference on the shared subband, the CPs of the participating operators can exchange compressed uplink baseband signals on finite-capacity backhaul links. This work tackles the problem of jointly optimizing bandwidth allocation, transmit power control and fronthaul compression strategies. In the optimization, we impose that the inter-operator privacy loss be limited by a given threshold value. An iterative algorithm is proposed to find a suboptimal solution based on the matrix fractional programming approach. Numerical results validate the advantages of the proposed optimized spectrum pooling scheme.

Cloud radio access network (C-RAN) is an emerging architecture for achieving high data rate and massive connectivity in future wireless communication systems. To further improve the performance of C-RAN downlink, this work considers the advantages of cooperation among user equipments (UEs). Specifically, it is assumed that the UEs can exchange the downlink received baseband signals via device-to-device (D2D) communication links which are orthogonal to downlink channel. The problem of jointly optimizing the C-RAN downlink and inter-UeD2d communication strategies is tackled with the goal of maximizing the sum-rate of UEs. To address the formulated problem, an iterative algorithm is proposed which attains a suboptimal solution of the problem. Numerical results confirm that the inter-UE D2D cooperation is effective to improve the downlink performance of C-RAN systems.

This work proposes an energy-efficient design of transmit power control and relaying strategies for the uplink of a cloud radio access network (C-RAN) with wireless fronthaul link. In the system, a set of single-antenna user equipments (UEs) send independent messages to a multi-antenna baseband processing unit (BBU) through a set of parallel multi-antenna remote radio heads (RRHs). The radio access link from the UEs to the RRHs is assumed to be orthogonal to the fronthaul link from the RRHs to the BBU. Since the overall performance of the system may be limited by the battery constraints of the UEs, this work tackles the problem of maximizing the worst energy efficiency of the UEs subject to the transmit power constraints at the UEs and the RRHs. Numerical results are provided to validate the advantages of the proposed energy-efficient scheme.

This work studies distributed uplink reception for a device-to-device (D2D) underlaid cloud radio access network (C-RAN) system, in which cellular user equipments (UEs) communicate with a baseband processing unit (BBU) through a set of distributed remote radio heads (RRHs) that are connected to the BBU via finite-capacity fronthaul links. The uplink of C-RAN system is interfered by a number of D2D communication links that operate in the same frequency band. Most of prior works on C-RAN uplink systems prescribe that the noise signals of different RRHs do not have statistical correlation. However, in the D2D underlaid C-RAN uplink systems, the effective noise signals that contain the D2D interference signals may have inter-RRH cross correlation. This work proposes an improved distributed reception scheme which leverages the noise correlation information. To tackle non-convex sum-rate maximization problem, a successive convex approximation (SCA) based iterative algorithm is derived that guarantees monotonically non-decreasing sum-rates. Some numerical results are presented that validate the advantages of the proposed scheme.

This work studies the uplink of a cloud radio access network (C-RAN) coexisting with a set of in-band device-to-device (D2D) communication links. In the system, the radio access communication from the user equipments (UEs), served by the C-RAN system, to the remote radio heads (RRHs) and the D2D communications take place at the same frequency band so that they interfere with each other while degrading the overall spectral efficiency performance. To mitigate this impact, the joint design of C-RAN uplink power control, D2D transmit power control as well as fronthaul quantization strategies is tackled with the goal of maximizing the sum-rate of all the UEs of the C-RAN and D2D communication systems. For the fronthaul quantization, both the Gaussian test channel and uniform scalar quantization approaches are considered. An iterative algorithm is derived based on the concave convex procedure (CCCP) approach to find an efficient solution, and numerical results are provided to examine the advantages of the proposed joint design as compared to the conventional separate optimization algorithms.