The Key Technology Of Cognitive Routing Protocol Research On Cognitive Wireless Sensor Network

Wireless sensor network (WSN) is one kind of dynamic wireless networks, which iscomposed of dense sensor nodes, and with high reliability and self-organization capability.These sensor nodes have the features of low-power, low-cost and small size, etc, and theycan communicate with each other by self-organizing. At present, as there are hugedemands of WSN, the amount of data that needed to be transmitted is getting bigger.Meanwhile, there indeed exists a huge gap between the routing management of WSN andthe big data management. Therefore, the WSN routing protocols has become a researchhotspot in both academia and industry.ZigBee is widely used in WSN for short distance wireless communication, which is alow-power, low-cost, short time delay and high reliability technology. This thesis mainlyintroduces the principles and the features of routing protocols in WSN, and focuses onZigBee routing protocol. ZigBee routing protocol is simulated under different networkenvironments on OPNET network simulation software. The simulation results show thatthe sensor nodes energy consumption can not be predicted in advance which is actuallyrandom. Thus, the energy consumption of each node is not average, which leads to somerouting problems such as low efficiency and long network delay.This thesis proposes a routing model based on reinforcement learning algorithm tosolve the problems mentioned above. Based on research work of reinforcement learningalgorithm, we introduce cognitive feature into the original routing protocol. The protocolshould perceive and evaluate the state of node energy, then modifiy routing strategyaccording to the value of nodes evaluation and the energy consumption. Solve the problemof node energy consumption and energy imbalance in the traditional routing protocol ofproblem, to optimize the network performance of the system requirements. This thesisuses OPNET network software as the simulation platform to verify and validate theperformance of cognitive routing protocol based on reinforcement learning algorithm. Bycomparing with the original ZigBee routing protocol, the protocol cognitive routingprotocol is verified and validated to be efficient enough to meet the requirements of WSN.