Polshchykova O.   Lazarev S.   Polshchykov K.   Konstantinov I.S.  

Model of neuro-fuzzy prediction of confirmation timeout in a mobile ad hoc network

Reporter: Polshchykov K.

Confirmation timeout (Round Trip Time, RTT) is an important value in data networks. Correct prediction RTT estimates the network load for the adequate selection of packet sending parameters and retransmissions parameters. In the TCP (Transmission Control Protocol) for the estimation of this quantity are used heuristic models. The vaues of coefficients in these models were obtained experimentally for a fixed topology networks. Therefore, the use of these models in the dynamic network (mobile ad hoc network) is inefficient.
The paper presents a prediction model of confirmation timeout, based on the application of fuzzy neural network theory. This model is based on the application of the Sugeno fuzzy inference algorithm zero order. The input values for fuzzy neural network are confirmation timeouts that measured in the current cycle and previous two cycles. The output value is the RTT which expected in the next cycle. The proposed model is set up and studied by simulations. In these experiments is simulated operation of a mobile ad hoc network that used for communication for opposition of emergency on dangerous construction sites.

The research have been carried out with the financial support of the Ministry of Education and Science of the Russian Federation (the unique identifier of the project is RFMEFI57815X0138).