Research On Efficient Algorithms For Attitude Measurement And Passive Tracking

According to different application scenes,localization technology can be divided into indoor localization with high-precision in a local space and satellite localization in a wide outdoor space.Indoor localization is to determine the position of people or objects in indoor scenes with existing cellular network or deployed sensor network.Satellite localization performs active navigation and passive non-cooperative tracking for objects on the surface of the earth by using navigation satellites or near-Earth reconnaissance satellites.For the above two scenes,this paper studies the efficient algorithm for localization and attitude measurement and the deterministic algorithm for passive tracking.(1)Rigid body localization(RBL)can be applied to helmets used in virtual reality(VR),indoor smart parking systems,docking of spacecraft,and position and attitude control of smart robots.In the scene of indoor localization,RBL system with single station based on the direction of arrival(DOA)is studied to realize the position and attitude estimation of the transformed rigid body in three-dimensional space.According to different fusion methods of observations and other known information,three maximum likelihood estimation(MLE)models are proposed.First,the transformed position coordinates of nodes are solved based on the distance between the node pairs fixed on the surface of the rigid body and the DOA measurement,and the rotation matrix and translation vector will be obtained by using the unit quaternion(UQ)method.Second,the rotation matrix and translation vector are solved according to the constraints of the rotation matrix and the DOA measurement.Third,the Euler angles of rotation and translation vector are solved according to the DOA measurement,and the rotation matrix will be obtained by the relationship between Euler angles and rotation matrix.The modified Gauss-Newton method is used to solve these nonlinear relations,and the rotation matrix and translation vector of the transformation can be obtained.The results of estimation are obtained through Monte Carlo simulation experiments and can reach the constrained Cramér-Rao bound(CCRB).The performance of the three methods is compared in the simulation.(2)Multi-satellite passive localization has the advantages of high location accuracy,large coverage,long range,and great concealment.It is widely used in navigation,surveying and mapping,rescue,electronic reconnaissance and other fields.Passive locating and tracking system with multiple satellites based on time difference of arrival(TDOA)and frequency difference of arrival(FDOA)is established to realize the position and velocity estimation of the moving source on the surface of the earth.Assuming that the errors in TDOA and FDOA are both unknown but bounded(UBB)errors,interval analysis is used to iteratively contract the initial position area of the source to obtain the position estimation interval.The least square is used to solve the initial area of the velocity that contracted by interval analysis similarly to obtain the velocity of the source.The true position and velocity of the moving source must be contained in the estimation interval.Simulation experiments demonstrate the effectiveness of the estimation results and also illustrates the influence of UBB errors on the estimation results.