WIPT receiver resource allocation strategy based on neural network

Simultaneously Wireless Information and Power Transfer (SWIPT) has been widely used as an emerging technology for energy-constrained networks. The existing SWIPT receiver resource allocation strategy only considers the optimal performance of the network at the current time. The network performance in the future is not considered, and the network with high quality of service requirements cannot be met. Aiming at this problem, a receiver resource allocation strategy based on neural network is proposed. The outage probability and its corresponding interrupt region in Delay-Limited (DL) transmission mode are considered from the perspective of energy and time. The resource allocation strategy that achieves the maximum throughput theoretically uses the neural network to predict the future time slot channel state information to dynamically adjust this strategy. The experimental results show that the proposed strategy can achieve more stable network performance under different channel conditions.