The Analysis And Simulation Of Distributed Multisensor Multitarget Tracking And Fusion System

The distributed structure of multisensor network, compared with the centralized structure, can decrease the bandwidth of data bus and the requirement of data process, and be easy to reconstruct the existed and decentralized multisensor system. Therefore, the distributed multisensor network has become the first choice while designing a new system. In this paper, a distributed multisensor multitarget tracking and fusion system is analyzed, the practice schemes of the key algorithms such as data associating and filtering algorithms, tracks associating and fusing algorithms and so on are studied, and the system performance is simulated and compared under different noise levels, different data associating and filtering algorithms and different tracks associating and fusing algorithms.In the system analysis, the simulated data detected by radar sensor including moving targets, fixing targets and faking targets dots caused by clutter are generated at first; Then the data are preprocessed so that the moving targets and shadiness dots are reserved, and the fixing and faking targets dots are deleted as more as possible; Furthermore, the reserved dots are processed to form local tracks using data associating and filtering algorithms; The space and time synchronization are realized through coordinate transformation and interpolation on time dimension; Finally, the local tracks are processed to form central tracks using tracks associating and fusing algorithms.In algorithms study, a hierarchy tree algorithm, which takes apart the validated matrix of the joint probability data association algorithm into many associated matrices, is put forward and optimized so that the optimized one is obtained; The joint probability data association algorithm is simplified, and the double probability choice algorithm is derived; The linear kalman filtering algorithm is extended under the mix coordinate of radar; The fuzzy double threshold...