A Real-Time Algorithm for Long Range Signal Strength Prediction in Wireless Networks

Xiaobo Long, Biplab Sikdar
2008 2008 IEEE Wireless Communications and Networking Conference  
Prediction of rapidly time variant fading channel conditions enables adaptive data transmission in wireless systems, which in turn improves the quality of service for end users and reduces the power consumption for data transmissions. In this paper, we construct an accurate, low-complexity, on-line prediction mechanism for the long range prediction of wireless link quality. Our method is independent of the propagation environment and distance between user nodes and the access points. The
more » ... d method uses past measurements of the received signal strength as its input, and uses a combination of segmentation, filtering and regression to predict the future trend in the received signal strength. An adaptive windowing mechanism is designed to adapt to abrupt changes in the data trace, which considerably reduces the prediction error. The algorithm is tested on real life networks in diverse environments. The prediction results are compared with one of the best existing channel prediction algorithm. We show that our algorithm can be used as a robust and comparatively more accurate predictor.
doi:10.1109/wcnc.2008.202 dblp:conf/wcnc/LongS08 fatcat:owxburdp5vhsjgp6gdyj6h44gq