JOINT SYMBOL LEVEL PRECODING AND RECEIVE BEAMFORMING FOR MULTIUSER MIMO DOWNLINK

While most of the existing symbol-level precoding (SLP) designs have assumed single-antenna users, this paper considers a joint SLP and receive beamforming (SLP-RBF) design for multiuser downlink. The formulated problem minimizes the total transmission power under the user symbol error probability (SEP) constraints. Due to the transmitter-receiver joint design, the problem involves non-convex bilinear terms and has a large number of SEP constraints. To overcome the challenges, we propose a new approximation method for the bilinear function that has an appealing separable structure. Based on the proposed approximation method, we develop a successive convex approximation algorithm for the joint SLPRBF problem, where the approximation problem can be efficiently solved by a dual based accelerated gradient projection method. Simulation results show the performance improvement of the joint SLP-RBF design over existing methods.