Abstract
A novel long-range prediction scheme is proposed for real-time motion picture expert group (MPEG) video to compensate the long-time delay in network traffic. Hence, it is useful for congestion control and dynamic bandwidth allocation to guarantee QoS at user-network interface (UNI). In this study, trend and periodicity of MPEG video are exploited in the design of prediction algorithm to improve the accuracy of prediction, especially for long-range prediction. The MPEG video sequence is represented by a state space dynamic, then Kalman and $H_{\infty}$ filterings are employed to predict video traffic output. Video traffic prediction can play an important role in dynamic bandwidth allocation and traffic management in future high-speed packet-switched networks and ATM networks. Unlike previous MPEG video traffic prediction schemes, which predict I, P and B frames separately, the proposed scheme predicts the composite MPEG video traffic levels. From several simulation results of real MPEG traffic data, the proposed schemes have a superior performance than the conventional methods in long-range prediction.