Research On Security Model Of Distributed Wireless Sensor Networks And Its Key Problems

As a new communication and data collection technology,distributed wireless sensor networks(WSNs)promote the unprecedented information revolution.With the characteristics of flexible deployment,high redundancy,low costs and easily feasible extension,distributed WSNs can be applied in the fields of multiple targets controlling,environment monitoring and the medical and military realms with high demands for security.Specially,in those regions that are out of reach such as hostile territories,extreme environments and depopulated zone,distributed WSNs are of great value for information collection.Due to the fact that security is the key to the widely use of distributed WSNs,the corresponding research achievements have been proposed and found in high-level international journals and conferences frequently.This dissertation analyzes and concludes the existing major problems in distributed WSNs according to the security of data,nodes and links.The topological structure and characteristics of distributed WSNs are analyzed in depth to clarify the security threats,which can be conductive to the construction of security architecture and reference model.Besides,the security schemes for distributed WSNs are presented depending on protocol designing,threshold calculating and security parameter analyzing.On this basis,the benefits of defense strategies could be optimized under the existing conditions,providing the basis for the applications of distributed WSNs in the fields of high security.The main research contents are as follows:1.Considering the problem that current researches are excessively focused on the schematic design for certain attack in distributed WSNs and the systematic study of security model is ignored,the security architecture and reference model are constructed according to the security requirements of distributed WSNs.Then the relationships between the nodes and security modules are clarified and the key problems still to be resolved are introduced in distributed WSNs.2.The traditional secure transmission mechanism is difficult to deal with the Rudolph attack and anonymous protection in distributed WSNs.Therefore,a trusted anonymous secure transmission model among clusters is introduced.In order to prevent the collusion between coordinator and certification organization,the negotiation mechanism of session keys is presented based on the trusted third party.Then the trusted anonymous authentication mechanism is designed in accordance with bilinear mapping,hashing function and parameters of elliptic curve.On this basis,the proposed model is realized by combining attribute set matching and anonymous authentication.Compared to other schemes,our scheme can effectively resist Rudolph attack and improves the efficiency of trusted certification.3.To overcome the problem of worm propagation in distributed WSNs,a modified epidemic model based description method for the states of nodes is introduced.It is proved that the ki-netic equations in the description method can create an accurate accounting of the number of nodes in different states.Then the optimal defense algorithm is put forward to minimize target cost function of the dynamic differential game between worm and distributed WSNs.In order to calculate the equilibrium solution,the existence of saddle-point solution is proved,which is solved by combining state parameters,cooperative variables and Hamiltonian functions.Experimental results show that the proposal can suppress worm propagation effectively and efficiently.4.Aiming at the problem that the performance of the communication deteriorates significantly in the presence of narrowband-interference(NI)attack in distributed WSNs,a novel algorithm for the detection of NI is proposed based on the frequency amplitude,which can explicitly calculate the threshold of transform domain according to the characteristics that the sub-bands are linearly independence and the noise follows the Gaussian distribution.In accordance with the utility function,modified CSMA/CA and regret value of the nodes,the multi-stage strategies for channel allocation are estimated iteratively.In the end,the effectiveness of the proposed algorithm is verified on the performance of NI detection and suppression.5.In the aspect of performance optimization of intrusion detection against various intrusion behaviors,this dissertation designs a dynamically multi-stage game model of intrusion detection in distributed WSNs.Based on the Bayesian rules and prior probability that external node is a malicious node in this stage,the posterior probability of external node and the set of node vulnerable to attack are formulated respectively in the proposed model.Then,the optimal defense strategy for intrusion detection is calculated accurately according to the conditions of perfect Bayesian equilibrium.Meanwhile,the analysis shows that developed scheme has distinct advantage in improving the success rate of detection and suppression in distributed WSNs.6.In order to solve the problem that partial cluster heads are drastically exhausted by the energy sink-hole attack in distributed WSNs,a novel suppression model based on the optimal path of mobile cluster heads is formulated,which is shown to balance the energy consumption in each cluster head.With the adoption of unequal cluster radius and data traffic monitoring explicitly,the inhibition mechanism is modeled as multivariate linear equations.Furthermore,toroidal projection is adopted to calculate the Euclidean distance,based on which the initial energy of mobile cluster head is configured correspondingly.In addition,the experimental results show that the developed algorithm can suppress energy sink-hole attack evidently and efficiently and is suitable to the distributed WSNs.In the end,a conclusion about the contributions proposed in this dissertation is summarised and the prospects for the future research direction are presented.