ACCESS FLOW-CONTROL SCHEME FOR ATM NETWORKS USING NEURAL-NETWORK-BASED TRAFFIC PREDICTION

The authors propose a new approach to the problem of congestion contro l arising at the user network interface (UNI) of ATM-based broadband n etworks. The access flow control mechanism operates on the principle o f feedback control. It uses a finite impulse response (FIR) neural net work to accurately predict the traffic arrival patterns. The predicted output in conjunction with the current queue information of the buffe r can be used as a measure of congestion. When the congestion level is reached, a control signal is generated to throttle the input arrival rate. The FIR multilayer perceptron model and its training algorithm a re discussed. Simulation results presented in the paper suggest that t he scheme provides a simple and efficient traffic management for ATM n etworks.