Prediction-based resource allocation for multimedia traffic over high-speed wireless networks

Both call admission control (CAC) and efficient scheduling are of great importance in next generation wireless networks, which are expected to handle various types of highly demanding multimedia users. In this paper, we present and evaluate a new mechanism which combines CAC with bandwidth allocation in a high-speed downlink time division multiple access (TDMA) wireless channel with errors; our mechanism incorporates predictions of the wireless channel condition in its decision making and our results show that, with the use of the feedback between the scheduler and the admission controller, system performance is significantly enhanced (in terms of voice-WAP-SMS-H.263 video QoS) compared to a scheme without prediction on the channel condition.