Jarkko Kaleva

Weighted sum rate maximization for interfering broadcast channel via successive convex approximation

We consider the weighted throughput maximization with general convex transmit power constraints in multi-cell multi-user multiple-input multiple-output system. The problem formulation is separated into receive beamformer and transmit precoder design problems. The non-convex precoder design problem is reformulated as a difference of convex functions program and solved with successive convex approximation. The convex approximation of the precoder design problem can be further formulated as a second-order cone program. Distributed and iterative solution via Karush-Kuhn-Tucker conditions for the precoder design with sum transmit power constraints is also proposed. It is shown that the rate of convergence can be significantly improved with the proposed algorithm while achieving comparable sum rate when compared to other recently published methods. This is an import factor in practical solutions as this increases the achievable sum rate with respect to required signaling iterations.

Downlink Precoder Design for Coordinated Regenerative Multi-User Relaying

Relays are used in the current wireless standards to enhance the cell throughput and coverage. In this paper, we evaluate the system performance by considering the duplexing loss with over-the-air transmission for relay with decode-and-forward as the relaying protocol. The overhead involved in providing the relay with the data prior to the transmission is often neglected. We demonstrate that the impact of relays can be even detrimental if the in-band signaling needs are not properly accounted for. Both precoder design and user allocation between the base station and the relay are considered. The precoder design objectives are transmit power minimization and throughput maximization each of which two different precoding approaches are considered. Zero-forcing precoder design is proposed to achieve reduced computation complexity and low signaling overhead. For higher efficiency and performance, we provide coordinated solutions, which allow more flexible interference control. We also consider user allocation by providing heuristic allocation methods.

Decentralized Sum Rate Maximization With QoS Constraints for Interfering Broadcast Channel Via Successive Convex Approximation

Weighted sum rate maximization (WSRMax) with user specific quality-of-service (QoS) constraints and general convex transmit power constraints is considered in multi-cell multi-user multiple-input multiple-output system. The particular focus in the proposed joint transmitter-receiver design is on tractability in terms of implementation and moderately fast changing/time correlated channel conditions. The non-convex transmit precoder design problem is formulated as a difference of convex functions program, for which a locally optimal solution is achieved by successive convex approximation (SCA). To achieve practically realizable designs, two decentralized approaches with low signaling overhead are proposed. Primal decomposition based solution provides better compliance of the provided QoS constraints in slowly fading channel conditions. On the other hand, solution based on Lagrangian relaxation of the coupling rate constraints is proposed for relaxed feasibility conditions and improved convergence properties. As a special case, an iterative solution via the Karush-Kuhn-Tucker conditions of the precoder design problem with per base station transmit power constraints is also proposed. Finally, we propose a heuristic extension of the SCA method, which is shown to significantly improve the rate of convergence while achieving comparable sum rate with recently published methods.

Bi-directional signaling for dynamic TDD with decentralized beamforming

This paper introduces a novel bi-directional signaling scheme for decentralized beamformer design to maximize the weighted sum rate (WSR) of multi-cell multi-user multiple-input multiple-output dynamic time division duplexing (TDD) system. In particular, the base stations are assumed to be either in uplink or downlink state based on the instantaneous traffic demand. The WSR maximization problem is solved via the corresponding weighted sum mean-squared error (MSE) minimization problem. An iterative algorithm is proposed with the bi-directional signaling embedded into the TDD frame structure. A detailed signaling architecture is proposed to facilitate channel estimation, exchanging the user beamformers and the user MSE weight matrices among the coordinating nodes. The proposed coordinated system is compared with the uncoordinated system while taking into account the overhead originated from bi-directional signaling. Finally, a periodic beamformer re-initialization strategy is introduced, which enhances the beamformer convergence rate and the system performance in a time-correlated fading environment. The proposed coordinated system provides significant performance gain in an interference limited environment.

Primal decomposition based decentralized weighted sum rate maximization with QoS constraints for interfering broadcast channel

We propose a decentralized beamforming design for weighted sum rate maximization with quality-of-service constraints based on primal decomposition of the corresponding centralized problem. The considered system consists of multiple coordinated multi-antenna base stations (BS) serving multiple users each with multiple receive antennas. The considered rate maximization problem is known to be NP-hard and as such intractable in practice. The problem is separated into downlink receive beamformer and transmit precoder design problems. The transmit precoder design problem is further decoupled between the interfering base stations by primal decomposition with respect to the caused inter-cell interference per BS. Particular emphasis is given to signaling requirements and rate of convergence.

Decentralized Linear Transceiver Design and Signaling Strategies for Sum Power Minimization in Multi-Cell MIMO Systems

This paper considers linear downlink transceiver design for the sum power minimization problem with per-user rate constraints in a multi-cell multi-user MIMO system. This non-convex problem is divided into transmit and receive beamforming optimization steps which are iteratively repeated such that the sum power converges. The transmit beamformers are solved using a successive convex approximation method (SCA), whereas the receive beamformers are computed via the linear minimum mean square error (MMSE) criterion. In addition to centralized design, we propose two decentralized algorithms where the transmit and receive beamforming designs are facilitated by a combination of pilot and backhaul signaling. In the first algorithm, sum power is minimized by computing the transmit beamformers at each base station (BS) and the receive beamformers at each user via iterative over-the-air optimization process. BS side processing requires local effective channel state information (CSI) acquired from precoded uplink pilots. In the second algorithm, the transmit and receive beamformers are iteratively optimized at each BS requiring only local CSI obtained from antenna-specific uplink pilots. In both algorithms, transmit beamforming optimization is decoupled among the BSs via primal decomposition method, which requires scalar backhaul information exchange for decentralized processing. The performance of the proposed algorithms is evaluated via numerical examples under different system settings in static and time-correlated channel conditions.

Decentralized beamforming for weighted sum rate maximization with rate constraints

A novel decentralized beamformer design for weighted sum rate maximization with user specific rate constraints is proposed for interfering multiple-input multiple-output (MIMO) broadcast channel (BC). In order to reduce the inherent complexity of the rate maximizing beamformer design, alternating receive and transmit beamformer generation together with successive convex approximation are applied. The rate maximization problem is reduced to mean-squared error (MSE) minimization with spatial data stream specific weighting factors that imply the inter-stream interference costs, where the weights depend additionally on the violation of the rate constraints. The iterative process is based on the Lagrangian relaxation of the rate constraints, and, thus, each intermediate beamformer solution is not required to be strictly feasible in terms of the rate constraints but rather steers the algorithm towards a solution with feasible rate constraints. Exchange of the weighting factors among the adjacent cells provides decentralized problem formulation for TDD based systems with a very low over-the-air signaling overhead.

User admission for multi-user regenerative relay MIMO systems

We propose an efficient user admission method for multi-user regenerative relay multiple-input multiple-output (MIMO) systems with maximum sum rate system performance objective. We extend the relay system model proposed in [1] to manage multi-antenna user terminals. Our aim is to incorporate computationally efficient user admission into the conventional weighted sum rate maximization. To increase the rate of convergence and mitigate over-the-air signaling requirements, we make early predictions of the active set of spatial data streams, which effectively translates to user allocation. It is shown that the proposed method achieves close to optimal user allocation and greatly improves the performance at low signal-to-interference-plus-noise (SINR) region when compared to direct channel quality based user admission [1].

Coordinated downlink precoder design for regenerative multi-user relaying

Regenerative relaying is an efficient method of increasing the achievable capacity and cell coverage. We propose a weighted mean squared error (WMSE) minimization based throughput maximizing precoder design subject to separate sum transmit power constraints for the base station (BS) and relay station (RS). The proposed algorithm is implemented in a distributed manner, thus, further improving the efficiency. We assume a system consisting of a single multi-antenna BS and multi-antenna RS, both of which are serving multiple single-antenna users. We consider a two-stage transmission strategy, in which, at the first stage, BS transmits to the associated users and the RS. At the second stage, the BS and RS transmit to their associated users. It is demonstrated that, with appropriate over-the-air (OTA) signaling, the system performance can be significantly improved.

Mode selection and transceiver design for rate maximization in underlay D2D MIMO systems

Linear coordinated transmit-receive beamforming methods are proposed for spatial underlay direct device-to-device (D2D) communication in cellular networks where both user terminals (UT) and base stations (BS) employ multiple antenna elements. For a D2D terminal pair, direct communication is a beneficial alternative compared to the cellular mode, where the direct data is relayed via BS. The transmitter and receiver design is formulated as a weighted sum-rate (WSR) maximization problem subject to individual transmit power constraints at BSs and users, where the utility is the aggregate sum of end-to-end data rates carried via uplink, downlink, and D2D links. The direct D2D links are allowed to co-exist with the cellular connections. The original problem is reformulated as an equivalent log-MSE minimization problem allowing alternating optimization of the transmit and receive beamformers. The precoder optimization step is efficiently solved using successive convex approximation (SCA) of the non-convex objective and constraints. According to the results, the system performance can be significantly improved with efficient mode selection between the BS and D2D transmission as opposed to solely serving all the users via D2D links or BS. Also, in certain scenarios, it is beneficial to allow a multi-route connection between the UTs.