Optimum Training Design for Wireless Energy and Information Transfer in MIMO Systems

• Main Authors: Chiao-Hsing Teng, 鄧喬馨
• Other Authors: Fan-Shuo Tseng
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/69237777969881900013

碩士 === 國立中山大學 === 通訊工程研究所 === 104 === In this paper, we study the joint precoder and channel estimation design for multiple-input multiple-output (MIMO) system with the energy harvesting (EH) device where there is one information transmitter (ITx) node, one information receiver (IRx), and one multiple antenna EH device. EH device can effect transfer the received radiation into energy at the receiver for extending the battery life. By the precoding techniques, we can further enhance the performance of the IRx and the EH simultaneously. However, the precoder design needs the downlink channel state information (CSI). Practically, the CSI can be estimated at the ITx with channel reciprocal (CR) property in time-division duplex (TDD) system. The longer training sequences lead to more accurate CSI but will condense the period of data transmission. Hence, there is a compromise between the accuracy of CSI and the information rate. In this paper, we adopt the minimum mean-squared error (MMSE) algorithm to estimate CSI and jointly design the length of the training sequence and the precoder so that the information rate and the harvest energy are optimized simultaneously. Since the optimization is a non-convex and mixed with integer programming, it is difficult to find the optimum solutions directly. We then propose an iterative approach to recursively optimize the training length and the precoder by two sub-problems. For the first sub-problem, we derive the precoder with fixed length of sequence and formulate the design as a convex formulation. The other sub-problem solves the optimum length of training sequence by the proposed bisection method. Numeral results show that our proposed design can effectively enhance the information rate under the EH constraint.