Distributed Underwater Acoustic Target Tracking Based On Random Finite Set Theory

Joint localization and tracking of underwater acoustic targets is one of the important problems in marine target detection,with significant applications in military operations,marine resource exploitation,and scientific research.Passive detection methods have advantages such as better concealment,lower cost,and lower power consumption compared to active methods.However,passive methods can not guarantee continuous reception of the acoustic signal from the targets,and the noisy environment with low signal-to-noise ratio increases uncertainty in obtaining targets existence information from measurements.The targets tracking algorithm based on random finite sets can identify targets existence from measurements and recover quickly after losing the targets,providing a suitable solution for joint localization and tracking of underwater acoustic targets.Additionally,with the development of distributed sensor network technology,sensor nodes can be placed in different spatial locations to cover the monitoring area,which can improve spatial gain and tracking performance.The background makes it possible to combine passive localization and tracking of underwater acoustic targets in complex environments where the number of targets is unknown and time-varying.This thesis focus on studying distributed underwater acoustic targets tracking based on random finite set theory.The main contents are as follows:Firstly,the thesis begins by studying the state-space targets tracking algorithm based on random finite sets,which provides the theoretical basis for the subsequent chapters.It discusses the target state model and measurement model in underwater acoustic environment,based on analyzing the evaluation indicators of tracking performance in this environment,the superiority of multi-targets tracking algorithm based on random finite sets regarding tracking accuracy and computational efficiency for complex scenes with an unknown and time-varying number of targets.Therefore,it is justified that the cardinality balanced multi-target multibernoulli filtering algorithm can be applied to filter measurements obtained via array signal processing in subsequent chapters.Secondly,the thesis proposed a underwater acoustic multi-targets joint localization and tracking method based on array signal processing.When localizing targets passively using array signal processing technology in a non-ideal environment,such as low SNR conditions,noise and interference can cause the main lobes corresponding to the real positions of the acoustic targets to be submerged by the peaks corresponding to the noise and interference and become sidelobes.To utilize the position information of acoustic targets that may exist in the sidelobes,two array signal processing methods,beamforming and matching field localization,are employed to generate measurements.The measurement design conforms to the characteristics of underwater acoustic physics,and the cardinality balanced multi-target multiBernoulli filtering algorithm is combined to filter out interference in the measurements,resulting in effective joint localization and tracking of underwater acoustic targets.Thirdly,the thesis proposed a distributed underwater acoustic multi-targets joint localization and tracking method based on array signal processing.By using a single hydrophone array as a distributed sensor node,longer detection range and higher resolution can be achieved while maintaining coherent signal processing gain for passive localization and tracking of non-cooperative targets.Furthermore,distributed sensor nodes can work together to fuse tarcking informationof the same target,resulting in higher tracking accuracy.To achieve this,we combine the cardinality balanced multi-target multi-Bernoulli filtering algorithm with the matching field localization and distributed fusion algorithm to form a distributed underwater acoustic multi-targets joint localization and tracking scheme.Our results show that this scheme achieves higher tracking accuracy compared to the underwater acoustic multi-targets tracking method under a single hydrophone array,Additionally,when compared to multi-targets tracking algorithm under centralized fusion structure,our method retains high tracking accuracy while sighificantly reducing the communication burden of the system and improving computing efficiency.Finally,the thesis proposed a distributed underwater acoustic targets joint localization and tracking method based on acoustic source of opportunity.The double-correlation function localization algorithm based on acoustic source of opportunity is found to be less sensitive to environmental mismatch than the traditional matching field localization algorithm.However,in non-ideal scenarios such as low SNR,the double-correlation function localization algorithm still generates sidelobes.Since these sidelobes may contain target position information with some probability,positions corresponding to peaks above the set threshold in the ambiguity function are used as measurements.These measurements are then fed into Bernoulli and CBMB filtering algorithm to obtain the targets trajectory in the monitoring area and improve targets state tracking results.This enables the realization of joint localization and tracking of distributed underwater acoustic targets based on acoustic source of opportunity.