Research On WSNs Cascading Failures In The Presence Of Malicious Nodes

Because wireless sensor nodes are susceptible to various factors,cascade failure of wireless sensor networks(WSNs)has become a hot issue for researchers at home and abroad.When the cascade occurs,the loss of packets by malicious dropping nodes can help to alleviate the cascade failure.Removing malicious nodes blindly is not conducive to the improvement of network robustness.It is necessary to explore the impact of malicious dropping nodes on the network in a more comprehensive way.In addition,the existing cascade failure optimization methods mostly start with single-step or local optimization to explore the anti-cascade optimization effect of the entire network,which is not conducive to the optimization of cascade failure resistance in the presence of malicious nodes.For the above problems,this paper takes low-rate WSNs as the background,studies WSNs cascade failure in the presence of malicious nodes,and designs a destructive optimization method for WSNs cascade failure in the presence of malicious nodes based on reinforcement learning,which is as follows.Firstly,the causes and operation process of cascade failure are analyzed,and the possible cascade events in various fields are listed.Based on WSNs network,cascade failure in the presence of malicious nodes is discussed,and the mitigation effect of malicious nodes on network cascade is deduced.It is pointed out that the current research on the positive impact of malicious nodes is inadequate and the existing cascade failure optimization schemes are not suitable for WSNs problems with malicious nodes.In order to provide theoretical basis for solving the above problems,WSNs networking protocol and intensive learning method are discussed.Secondly,for the overall impact of malicious nodes,a cascade failure-resistance model of WSNs is established,which takes the actual amount of packets sent and received by sensor nodes as load.The impact of malicious nodes is fused into the model,and the rationality of the model is analyzed and verified.Based on this model,a real WSNs network is constructed,and the WSNs cascade failure phenomenon in the presence of malicious nodes is analyzed experimentally.The results show that the packet loss behavior of malicious nodes can greatly alleviate the cascading failure of WSNs under certain conditions;And different load distribution modes show different cascade characteristics;In addition,in most low-speed WSNs,preserving malicious nodes during cascading can improve the invulnerability of the network.Finally,since the existing methods are not suitable for cascade failure-resistance optimization in the presence of malicious nodes,this paper uses the reinforcement learning algorithm to design a cascade failure-resistance optimization method for WSNs based on Q-Learning,and analyzes the impact of the relevant parameters based on this method.The simulation results show that the method can achieve the cascade results matching the objective optimal solution.The load carried by the attacking nodes and the setting of the maximum number of iterations in simulation have an impact on the simulation results.Malicious packet loss nodes can get higher returns if they exist in the cascade network.