Research On Key Technologies Of Target Nodes Localization In Ocean Sensor Networks

The countries across the world have drawn much attention to the ocean after the improved understanding of marine resources exploration,the scientific research value of marine resources,and the strategic position,which leads to the birth of the slogan for vigorously developing the marine economy.In this context,China also proposes the strategic goal of "building a maritime power," which has been mentioned in many reports,so-called "persist in land and sea coordination,develop the marine economy,and build a maritime power." As one of the necessary supports for the development of the marine economy,ocean monitoring technology based on ocean sensor networks(OSNs)is conducive to improving marine public service capabilities and environmental perception capability for intelligent vessels,including disaster risk assessment based on environmental perception data and state perception and prediction based on necessary marine information.Besides,OSNs can provide a better technical measure and information platform for safety assurance,marine environmental protection,and safety navigation for intelligent vessels.In what concerns OSNs,the data collected is meaningless if they are not geo-referenced.Therefore,how to acquire relatively accurate location information that we call localization is a requirement for marine monitoring applications and the basis for the other related issues such as routing and topology of OSNs,or vice versa.However,it is quite challenging to obtain relatively accurate location information for the target node in OSNs with the complex and changeable ocean environment.On the one hand,for underwater nodes that cannot receive Global Positioning System(GPS)signals,it is common to use nodes with GPS at the sea surface to assist in localization via the underwater acoustic signals.Nevertheless,adverse effects of the attenuation and absorption loss on acoustic signal propagation in the inhomogeneous medium may degrade the localization accuracy.On the other hand,the corresponding deployment cost would increase if all anchor nodes deployed at the sea surface are buoys with GPS information.As an alternative,a common strategy is to use a part of the buoys with GPS information to locate the other nodes at the sea surface,through which the nodes without GPS information can be the reference nodes to assist in localization for the underwater node.However,for nodes that communicate through radio signals at the sea surface,negative factors like the high dynamics of the marine environment,the non-linear and non-Gaussian noise generated by the shadowing effect of ocean waves,and the uncertainty in transmission power of nodes and environmental model parameters may increase the localization error as well.To this end,the dissertation conducts systematic research from two aspects for localization in OSNs,i.e.,localization for the surface node and localization for the underwater node,according to the problems mentioned above.The main research content of the dissertation includes:1)In order to improve the localization accuracy of the underwater target node,an information-driven OSNs optimal placement strategy is proposed.A closed-form of Fisher information matrix(FIM)is conducted after using the particle-based theory to quantize the hybrid measurement noise(the attenuation noise and absorption noise).Then,the objective function based on FIM is constructed,which is solved by the A-criterion after a relaxation operation.An optimal placement strategy is presented further with a specific condition through the solution of the objective.The simulation experiments demonstrate the effectiveness of the proposed placement strategy and,based on which,optimal placement of the surface nodes for underwater target node localization is illustrated.2)For the relatively low accuracy of localization caused by ocean environment dynamics and the deviation between the rated and actual transmission power,a localization method with the uncertainty in node transmission power in OSNs is proposed.The original non-linear and non-convex localization problem is transformed into an optimization problem under the alternating nonnegative constrained least squares(ANCLS)framework.Then an active set method(ASM)is utilized to figure out the feasible solution with the exchange of the solution index between the active set and the passive set.Since it is easy to drop in the local optimum,a majorization-minimization strategy is used to estimate the transmission power and location jointly.3)A joint estimation method for the path loss exponent and location information in OSNs is proposed to mitigate the localization error caused by the uncertainty in environmental model parameters.The optimization function is conducted after the first-order Taylor series expansion for three times together with the relevant regularization parameters and robust function.The problem is then divided into two sub-problems,which are solved through iteration with a block coordinate update(BCU)method.4)An improved particle filter(PF)cooperative localization approach for multiple targets in OSNs is proposed with the cooperative theory,with which the likelihood function of the system is modified with an optimal parameter obtained by minimizing the relative entropy.The location of the target node is refined via iteration.According to the method,the problem of the inefficient and low coverage on localization is improved when existing multiple target nodes.5)An absorption mitigation localization technique in OSNs is presented to improve the localization accuracy caused by the absorption loss over the rise in transmission frequency.The localization problem is converted into an optimization problem via the first-order Taylor series expansion.After combining with the constraints,an improved generalized trust region sub-problem(GTRS)framework is then conducted with a min-max operation.The solution is acquired via iteration,in which the adverse impact of absorption loss on localization is mitigated.The research results would help to improve the positioning efficiency,accuracy and robustness of OSNs in the face of various adverse conditions,provide environmental data with more accurate location information for the safe navigation of intelligent ships,and promote the ability of environmental perception.Meanwhile,the proposed methods have positive guiding significance for promoting the application of wireless sensor network positioning in other similar harsh environments,and also have important scientific significance and academic value for enriching the positioning related theories.