Link Quality Prediction Using Shadowing Time-Frequency Correlation in Multi-Carrier Wireless Networks

This paper addresses the problem of link quality prediction on a secondary frequency carrier from measurements on a primary carrier during a scanning interval in multi-carrier wireless networks. An MMSE linear predictor is proposed which makes use of the information available in particular on shadowing time and frequency correlation. Compared to state of the art, the proposed solution achieves a significant gain in accuracy of prediction, up to 2.2 dB in average RMSE for 30 kmh mobile speed and IEEE 802.16e maximum scanning interval duration of 1.275 sec. The accuracy gain decreases when the time increases or the mobile speed increases, but it always keeps positive, thus proving the superiority of the proposed solution as compared to state of the art.