Traffic Prediction Based Energy Efficient Transfer for Point-to-Point Wireless Communications

This paper investigates the performance of Kalman filtering based energy-efficient transfer for nonlinear stochastic discrete-time systems and proposes a low complex traffic load prediction algorithm. Based on the historical values of traffic load, the transmitter estimates the traffic load of the current time slot. And with the predicted traffic load, the transmission power in the incoming time slot is adjusted by the transmitter adaptively. Simulations are conducted to validate the theoretical analysis results, the proposed traffic load prediction algorithm can improve the energy efficiency of the wireless networks. The prediction accuracy is also validated with the real data sets of a local base station.