Research On Distributed Secure Localization Issue Of Wireless Sensor Networks

With the continuous advancement of sensing technology,wireless communication and embedded computing,as well as the vigorous development of digital circuits and MEMS,as a distributed information system that integrates data acquisition,storage,transmission and control processing,WSN is capable of real-time perception and monitoring information about the environment or targets in the deployment area.As a key technology of WSN,localization can achieve various functions such as spatial perception,target tracking,environmental monitoring,and emergency rescue by providing location information.High precision localization can enhance the network efficiency,improve the effectiveness of data association and decision support,hence it is crucial for the practical applications of WSNs.However,due to the openness of wireless communication,the uncertainty in deployment environments,and the constraints imposed by limited resources,WSNs are highly vulnerable to various security threats,resulting in localization inaccuracy,which in turn deteriorates system performance or even leads to system crash.Consequently,in recent years,the issue of secure localization of WSN has gradually gained the attention of researchers.The outcomes of the research have involved multiple fields,including wireless communication,network security,and coordination control,but there are still certain limitations.First of all,most of the existing research on secure localization is based on range-free methods.Due to the adverse effects of environmental changes,signal propagation loss,and other factors,its positioning accuracy is relatively low.It is necessary to explore secure localization schemes based on range-based technology.Secondly,most existing security localization algorithms are usually designed for a specific form of attacks,which limits their applicability in resisting different attacks.There is an urgent need to establish a universal attack model to design more robust security localization algorithms.Thirdly,current security localization research mainly focuses on the malicious attacks on nodes,but neglects the potential threats to wireless communication links.It is required to concern the communication link security,so as to provide strong guidance for developing more universal security localization methods.Furthermore,ignoring the privacy preservation of node locations may bring serious security threats,while traditional encryption and authentication mechanisms often incur large computational overhead.Studying how to efficiently achieve localization in WSN while preserving location privacy is an urgent scientific problem to be solved.In view of the issues existing in the aforementioned research status quo,combined with the actual situation of the development in the new generation of WSN,this dissertation aims to address the challenges of distributed security localization of WSNs from the four key perspectives of data integrity,data availability,link validity and data confidentiality.The specific research contents are summarised as follows:1.In response to the issue of existing range-based localization methods are difficult to defend against threatening deception attacks,based on the analysis of the fundamental principles of localization and the models of typical security threats in WSN,the mathematical models of three specific forms of deception attacks involving the manipulation of distance measurement values based on RSSI are formulated at first.Secondly,to resist the surge injection attack with the goal of maximizing instantaneous disruption,a trust assessment scheme is introduced for triangulation sets based on the statistical characteristics of RSSI distance measurements.This is coupled with mean filtering and data fusion techniques,leading to the design of a distributed localization algorithm based on trust assessment strategies.Furthermore,in dealing with cumulative bias injection attacks,a fixed time interval detection scheme is proposed.Building on the outcomes of anomaly detection,and incorporating a numerical fitting method,a distributed localization algorithm based on a bias correction strategy is developed.Simulation results demonstrate that the aforementioned algorithms can achieve accurate localization in the presence of deception attacks in which distance measurements have been tampered with.2.In view of the problem that the range of attack types that can be defended against by the current security localization research is too specific,combined with the features of DoS attacks and the practicality of network applications,a universal model of periodic DoS attacks is characterized.In this model,each attack consists of an active period and a dormant period.There is no need to make any assumptions about the intrinsic strategy of the attacks,and the behaviors of attackers are only determined by the duration and frequency of the attack.The model allows for a flexible trade-off between efficiency and resources.Subsequently,on the basis of the framework of a distributed iterative localization scheme based on barycentric coordinate representation,an update rule based on a hold-on strategy is introduced,which allows the sensor to update its coordinates using packets collected from its neighbours at the last moment of the previous attack dormant period when subject to an attack.Furthermore,theoretical proofs are presented to demonstrate the algorithm’s convergence to the accurate locations of sensors without taking the attack strategies into account.Finally,experiments were conducted on the platform of Raspberry Pis,with results aligning with the theoretical findings.3.In regard to the prevalent oversight of communication link security in existing research,firstly,a more generalized model of random lossy links is proposed based on a Bernoulli random variable to characterize the effectiveness of wireless communication links of sensor nodes from the controller to the actuator.Secondly,barycentric coordinates are determined by measuring the distances between pairs of nodes in the WSN.These coordinates are employed as the basis for a range-based localization scheme,and a more universally applicable distributed iterative localization algorithm is proposed.Subsequently,by leveraging the convergence properties of sub-stochastic matrix products,the global convergence of the proposed algorithm is theoretically demonstrated.The results of complexity analysis show that the algorithm entails low computational and communication overheads.Finally,the accurate localization performance of the proposed algorithm subject to random lossy links is verified through numerical examples.4.As for the contradiction between privacy preservation and resource constraints in current research,considering the issue of time-varying communication delays in practical scenarios and through a comparative analysis of prevalent privacy preserving methods such as encryption,state decomposition and differential privacy,a theoretical model is established for the problem of distributed privacy-preserving localization at first.Secondly,a distributed localization algorithm is proposed based on a dual privacy-preserving strategy involving asynchronous iteration and noise injection.The asynchronous iteration strategy not only avoids privacy leakage,but also reduces the communication overhead and mitigates the risk of network congestion.The injected noise can be completely cancelled out during the localization process,thus the localization performance of the system will not be degraded.Furthermore,the effectiveness of the proposed privacy-preserving localization algorithm is verified from three aspects:algorithm convergence,complexity,and security efficiency.Finally,the demonstration is further carried out through experiments on an unmanned vehicle platform,affirming that the proposed algorithm achieves accurate localization while upholding the confidentiality of node location information.