In order to inprove the tracking performances of some sensors and local nodes

the paper discusses the state estimation techniques in multilevel multisensor surveillance systems with feedback information.Based on the single sensor Kalman fitering equations with feedback information

this paper presents two-level centralized

distributed and hybrid track level fusion methods with feedback information in multicoordinate systems.In the different Cartesian coordinates

several kinds of track level fusion methods with feedback information for three-level multisensor data fusion systems are proposed

in which centralized-distributed

distrbuted-distributed and hybrid-distributed estimation combination problems with feedback information are considered

and proves that two kinds of three-level estimation solutions with or without feedback information are optimal and equivalent in the form of the theorem.The simulation results show that the multilevel multisensor data fusion system can not only increase the global estimation acuracy of target but also greatly improve the tracking perfomances of some sensors and local nodes by using the feedback information.