Near-Field Wideband SWIPT With Delay-Phase Precoding: Beam Training and Optimization

Abstract

To address the dual demands of high data transmission and energy supply in sixth-generation (6G) networks, this paper investigates near-field wideband simultaneous wireless information and power transfer (SWIPT) with power splitting for multiple mobile stations. To mitigate the beam splitting effect, we introduce fully- and sub-connected delay-phase precoding (DPP) architectures. Under energy harvesting constraints, we formulate a sum spectral efficiency maximization problem by jointly optimizing phase shifter (PS)- and true-time-delay (TTD)-based analog precoders, digital precoders, power allocation, and power splitting factors. We reveal controllable beam splitting and beam focusing phenomena in near-field wideband DPP, which can be flexibly adjusted through TTD and PS tuning. For the fully-connected DPP, we propose an efficient beam training scheme via frequency-domain angle search and time-domain distance search. Moreover, for the sub-connected DPP, we develop a structured precoding approximation matching (SPAM) mechanism. Based on beam training results, we design information decoding-oriented and energy harvesting-oriented digital precoders to balance the spectral efficiency-energy harvesting trade-off. Finally, we propose an alternating optimization framework and a low-complexity sequential quadratic programming method to solve the joint power allocation and splitting optimization. Simulation results demonstrate that the proposed schemes achieve superior spectral and energy efficiency in near-field wideband SWIPT systems.