Research On Localization Algorithms And Security For Indoor Wireless Sensor Networks

Wireless sensor network,a self-organizing network consisting of a large number of sensor nodes,has been widely applied in a broad range of fields such as military,environmental monitoring,medical care and smart home.The study of indoor localization technologies for wireless sensor network has recently attracted great attention because of the increasing demand for indoor location-aware applications.Many indoor applications of wireless sensor networks require precise knowledge of the locations of sensor nodes.Designing indoor localization algorithms with high-accuracy,low power and high security has recently been one of the hottest research topics in wireless sensor network.The localization algorithms for indoor wireless sensor network are studied in this thesis,and the main contributions are:(1)The propagation characteristics of wireless signal for wireless sensor network in indoor environments are summarized by analyzing experiments conducted in office,hall and underground tunnel environments.The mainly influential factors are identified for indoor node localization.(2)Different transmission power can cause serious the accuracy degrade of indoor positioning.To avoid this problem,a robust linear localization algorithm called VTPLL is proposed.The algorithm adopts received signal strength indicator gathered from real indoor environment.We include the transmitting power as a critical parameter,which increase the flexibility in nonlinear optimization in localization.We transform the traditionally nonlinear localization problem into a linear optimization problem using a linear approximation method.Simulations and experiments show the outperformance of the VTPLL algorithm compared to existing algorithms.(3)The accuracy of RSSI-based localization is severely affected by multipath effect and external radio interference in underground tunnel environments.To solve this problem,we propose a novel RSSI-based localization algorithm which overcomes the multipath effect by applying geometrical optics rules in analyzing RSSI measurements in both Line-of Sight and Non-Line-of-Sight scenario.The proposed algorithm can reconstruct a function for path loss and the physical distance which can help minimize the localization error caused by multipath transport.We conduct extensive real-world experiments in underground tunnel environments and analyze the accuracy of localization of the proposed algorithm.We further discuss the impact of some representative system parameters on ranging,including multipath attenuation factor and path loss exponent.(4)A new decentralized gradient descent localization algorithm is proposed in order to reduce the communication overhead and the funneling effect in indoor wireless sensor networks.The proposed algorithm is based on gradient descent and can be carried out in a distributed and iterative manner.In each iteration,the position of a target node is calculated by combining a weighted average of the previous position estimate and the local observation.Simulation results demonstrate that the proposed algorithm performs better than existing algorithms.(5)Localization of sensor nodes often suffers from attacks launched by adversary that can distort transmitting power and degrade localization accuracy in wireless sensor network.To solve this problem,we propose a robust security localization algorithm based on radio measurements.We present two security models that address single target and multiple targets,respectively.We transform the nonlinear localization problem into semidefinite programming problem.Simulations and Experimental results demonstrate that the proposed algorithm outperform existing algorithms.(6)First,the shortcoming of existing secure localization algorithm in wireless sensor network is analyzed.Then,a new secure localization algorithm based on Newton method is provided,meanwhile an anomaly detection is studied and improved.The performance of the algorithm is evaluated by a combination of simulations and experiments which prove that WSN nodes running this algorithm are able to locate themselves in hostile or untrusted environment with higher precision and better stability.