Incorporating Target Mensuration System for Target Motion Estimation Along a Road Using Asynchronous Filter

In support of TNT experiments, the NPS UAV laboratory has developed a Vision-Based Target Tracking (VBTT) system for a Small Unmanned Aerial Vehicle (SUAV). This system provides an autonomous target tracking capability, while simultaneously estimating the target’s velocity and position. The accuracy of the existing system can be improved by providing external corrections to the target position estimation from the geo-rectification system (GIS). This thesis addresses the implementation of an asynchronous correction scheme into the target position estimation filter. The current autonomous position estimation algorithm provides 20-30 meters accuracy. The external correction system (Perspective View Nascent Technologies (PVNT)) is expected to provide target position accuracy of 1-2 m. However, a delay of up to 10 seconds is expected. Therefore, in order to improve the accuracy of current estimation of target motion, a new asynchronous correction technique that incorporates the more accurate PVNT data is proposed. To further improve the target motion estimation, it was also proposed to incorporate a known road model into the filter and compare its performance with the original filter.