Missing trajectory estimation using single station EOT system data

During test and evaluation (T&E) process of airborne objects/targets, electro-optical tracking (EOT) system has garnered importance for providing accurate and precise positional information of trajectory along with monitoring of various critical events. In absence of a ranging sensor like LRF (Laser Range Finder) or RoR (Range only Radar), a single EOT system only provides line-of-sight (LoS) information i.e. azimuth and elevation as captured through its passive optical (IR & CCD) sensors. However, range information is obtained with the LOS data of two or more time-synchronized EOT systems, positioned at distributed geographical locations (for a fairrelative apex angle of LOS),by means of triangulation in order to furnish the positional information with respect to time. It is pertinent that EOT system is likely to miss the target of interest (ToI) due to various environmental constraints. Hence, during some T&E processes, it is not feasible to produce a portion of trajectory information even with the data available from a single EOT system. This paper aims towards extracting data of such missing portions, through a novel approach, based on estimation correlated with LOS data available from a single EOT system. The approach has also been incorporated through a neural network model that estimates the missing segment accurately under supervised learning. Extensive offline experiments have been conducted to assess the efficacy of the proposed prediction paradigm with the results of remarkable improvement.