An Adaptive Sampling VB-IMM Based on ADS-B for TCAS Data Fusion with Benefit Analysis

摘要

This paper discusses the problem of data fusion of AutomaticrnDependent Surveillance Broadcast and Traffic Alert and CollisionrnAvoidance System. First, the 3-dimensional trajectory is generated byrnaircraft movement model. Contrasting to traditional aircraft surveillancernresearch focusing on data precision with known noise and fixed samplingrnperiod, the estimation of time-varying noise is guaranteed by VariationalrnBayesian method, which is the basis for failure prediction and adjustment ofrnsampling period. Second, the interacting multiple model is used for localrnfiltering. Two situations are considered during fusion, including scenariosrnbefore and after injecting Automatic Dependent Surveillance Broadcastrnfailure modes. Then, data fusion’s benefit for improving the false alarm andrnleak alarm is analyzed by calculating the time until closest point ofrnapproach between aircraft. Finally, simulations results are given to verifyrnthe validity of the algorithm proposed in this paper. It shows that therndynamic noise can be estimated within a tolerable error range and thernsampling period can be adjusted according to the noise level. Compared tornsingle Traffic Alert and Collision Avoidance System, Automatic DependentrnSurveillance Broadcast system and current statistical model based fusionrnsystem, the root mean squared error, alarm condition can be optimized andrnfailure at information level can be detected.