SWIPT for MISO Wiretap Networks: Channel Uncertainties and Nonlinear Energy Harvesting Features

This paper investigates the power minimization problem for a simultaneous wireless information and power transfer (SWIPT) system in MISO wiretap networks, where one multiple-antenna transmitter intends to transmit required amount of information and energy to its legitimate receiver while restrict the information leakage to an eavesdropper (Eve). The nonlinear EH model is employed for SWIPT. Two uncertainty MISO channel models are considered for the legitimate receiver, i.e. the deterministic uncertainty model (DUM) and the stochastic uncertainty model (SUM), and the Eve is assumed not to feed back its channel to the transmitter. For the DUM, the worst-case design with global optimum is solved by our proposed method based on semidefinite relaxation (SDR) and S-procedure. For the SUM, the statistically robust design with a tight upper bound to global optimum is obtained by our proposed method based on SDR and Bernstein-type inequality. Numerous simulation results demonstrate the validity and efficiency of our proposed robust transmit design methods. Compared with the traditional linear EH model, employing the nonlinear EH model can avoid false output power at the legitimate receiver or save power consumption at the transmitter as the real circuits are working in the nonlinear output field rather than the linear one.