Optimization for Multicarrier MIMO SWIPT Systems Under MSE QoS Constraint
Autor: | Minghua Xia, Xingxiang Peng, Peiran Wu |
---|---|
Rok vydání: | 2020 |
Předmět: |
Semidefinite programming
Mathematical optimization Optimization problem Computer science MIMO MathematicsofComputing_NUMERICALANALYSIS 020206 networking & telecommunications 020302 automobile design & engineering 02 engineering and technology Transmitter power output Precoding Matrix (mathematics) 0203 mechanical engineering Convex optimization 0202 electrical engineering electronic engineering information engineering Interior point method Computer Science::Information Theory |
Zdroj: | VTC Spring |
DOI: | 10.1109/vtc2020-spring48590.2020.9129120 |
Popis: | This paper studies the joint transceiver design and receive power splitting (PS) optimization for a multicarrier multiple-input and multiple-output (MIMO) simultaneous wireless information and power transfer (SWIPT) system. We aim to maximize the harvested power at the receiver under a sum mean-square error (MSE) constraint and a total transmit power constraint. By first deriving the linear minimum MSE equalization matrices, we obtain a non-convex optimization problem involving the transmit precoding matrices and the receive PS ratio. For a given PS ratio, we show that the problem can be recast as a semidefinite programming (SDP) problem, which can be solved with the interior point algorithm. Then, by employing the unimodal property of the objective function with respect to the PS ratio, we propose a Golden-section search method to find the optimal PS ratio efficiently. Further, to reduce the complexity of solving the inner SDP problems, we exploit the optimal structure of the precoding matrices to transform the original matrix-based optimization problem into a scalar-based optimization problem, for which the closed-form solution is obtained through convex optimization techniques. Simulations are provided to validate the superior performance of our proposed solutions. |
Databáze: | OpenAIRE |
Externí odkaz: |