Kalman Filter Tracking of a Reflective Target Using Forward Looking Infrared andDoppler Return Measurements

摘要

Plume phenomenology experiments conducted in 1990 uncovered the existence ofplume speckle reflectance emanating from the exhaust of a solid-propellant rocket motor due to the presence of metallic particulates in the plume. This impeded attempts, that are dependent upon the speckle return of an actively illuminated target hardbody, to discern the plume/hardbody interface. This thesis investigates the feasibility of employing the Doppler return phenomenon in discerning the plume/hardbody interface and locating the hardbody center-of-mass. The potential of the Doppler phenomenon's utilization lies in the fact that the Doppler return frequency spectra of the plume and the hardbody respectively possess distinct properties and are differentiable from each other. Hence, these differences can be exploited in the attempt to discern the plume/hardbody interface. For this study, two center-of-mass Kalman filters are developed to receive the Doppler return measurements: a one-state filter and two-state Modified Maximum a Posteriori Multiple Model Adaptive filter. A sensitivity analysis is conducted wherein the performances of the one-state filter and two-state filter are evaluated with variations in transmitted wavelength, signal-to-noise (SNR), and probability-of-miss. Results show the center-of-mass filters are sensitive to increases in probability-of-miss, whereas decreases in SNR produced insignificant degradation in performance. The two-state Modified MAPMMAF achieved the best performance, and clearly has the potential to accomplish the task of locating and tracking the hardbody center-of-mass. (Author)