K.A. Elbatsh, E. Macas Lpez, and A. Surez Sarmiento (Spain)
Handover prediction, RSSI gradient, RSSI filters, Kalman filter, Grey model
Physical properties of wireless channels make very difficult or sometimes impossible to maintain open client sessions for interactive multimedia services. Short term client sporadic and not previsible disconnections lead to service interruptions which provoke a user headache especially if a firm real time multimedia service is used and a hard error when a real-time service (e.g. vigilance) is used. Also, during handover process channel conditions deteriorate leading to data loss. With multimedia services, handover influences packet loss, this is avoided by handover prediction mainly. Many handover prediction strategies use Received Signal Strength Indicator (RSSI) as the basic indicator for disconnection. Filters like Kalman and Grey Model (GM) are used by many researchers to estimate the RSSI value. These filters adapt rapidly the prediction to the continuous RSSI. This means that they can not work with short disconnections. In this paper we present a new RSSI Gradient predictor and filter technique capable of controlling short disconnections efficiently. The RSSI Gradient filter has been theoretically tested supporting well a long set of random Mobile Client (MC) movements inside the coverage area where coverage holes can be considered novelty.
Important Links:
Go Back
