Data Fusion algorithms to improve test range sensors accuracy and precision

Data fusion technology is used in many research areas, such as image processing, automotive, robotics, defense and aerospace. The need to fuse data from different sensors reflects the demand for enhanced measurements precision. Every sensor has pro and cons, data fusion technology for tracking purposes in a test range is supposed to leverage the highest quality features of each sensor to obtain a better position estimation. This paper presents NT multi-sensor tracking algorithm, it is developed to receive target position information from different tracking sensors, radar and electro-optical, and use data fusion technology to obtain better quality data compared to the one provided by a single sensor alone. The algorithm uses a fusion technique which permits a huge performance boost: it associates a weight to every sensor calculated using Kalman filters, properly modeled, to estimate sensor measurements deviations from the estimated target position. These weights help the algorithm understanding which sensor is more reliable and to estimate the best possible position, which is very close to the real target position. Real case tracking scenarios are presented, where the algorithm is tested and compared with reference trajectory data to estimate its accuracy.