Object Tracking with Missing Data Using Predictive and Filtering Methods

Achieving high precision in object tracking has become a essential objective across various sectors, encompassing industry, military, civil, and beyond. During object tracking, information loss may occur in the data acquisition phase due to measurement errors or partial or total occlusion of the object, resulting in missing segments of the trajectory and therefore, larger tracking errors. Issues like bounded sensors, environmental disturbances, and communication limitations frequently distort measurement features, compromising the integrity of available data. This paper proposes leveraging predictive algorithms integrated with GUFIR and GKF filtering algorithms to address lost measurements or missing data during object tracking under colored measurement noise (CMN). The performance of GUIFR predictive was found to be superior to GKF predictive, with both algorithms significantly outperforming accuracy with measurements under CMN and missing data. In our research, we present a promising approach to enhance object tracking accuracy under CMN and missing data, offering valuable insights for future developments in tracking algorithms.