Modeling And Performance Analysis Of DCF Protocol In Wireless Sensor Network

As a new wireless self-organization network, wireless sensor network has beenwidely used in military defense, environmental science, health care and so on. Modelingand performance analyzing for wireless sensor network has been a hot spot in currentwireless network research.IEEE802.11 as the primary standard of wireless network communication protocol,defines Medium Access Control (MAC) layer, its effect is to distribute the limitedbandwidth of wireless channel resource for nodes effectively, and the availability of theMAC protocol is the main factor that affects the performance of the system. In this paper,two analytical models based on collision probability are proposed for the binaryexponential back-off (BEB) algorithm which is used by the IEEE802.11 DCF mechanism.Firstly, the saturation performance analysis model of the network based on thecollision probability is obtained, by considering the influence of the number of back-offand number of retransmission in the binary exponential back-off mechanismcomprehensive, using a method of probability average to obtain a simplified model of theBEB algorithm, based on the model, the relational expressions of the collision probabilityp between the system throughput and delay are established, analysis its performance underdifferent number of nodes and the length of data packet, and then compute the handoffthresholds of the two access mechanism (basic access mechanism and RTS/CTS accessmechanism) under different number of nodes.Secondly, the non-saturated performance analysis model of the network based on thecollision probability is obtained, by considering the influence of the number of terminals,traffic loads, number of back-off and number of retransmission in the binary exponentialback-off mechanism comprehensive, use the M/G/1/K queuing system to reflect the sendqueue of each terminal free or not to show the saturation and un-saturation of network,and used the method of probability average that used in the first model to evaluate theaverage service time of data packet, and then use the Bianchi model, adding another stateX based on it to model each state of the terminal, the normalized throughput and delay are evaluated, then analysis its performance under different number of nodes and packetarrival rates.Lastly, build the simulation network use OPNET, numerical results obtained by themodel and simulation results were compared to verify the accuracy of the models, and itprovides a theory basis for improving the agreement and optimizing the performance ofthe system.