Minimum variance missile launch and impact estimation by fusing observations from multiple sensors

Using minimum variance estimation techniques, we are developing a model-based, post-event missile trajectory and error analysis program called BOOSTR. The program´s development has been motivated by a requirement to determine the errors in launch point estimates, and how these errors are affected by fusing data of various types from multiple sensors. BOOSTR´s error analysis capabilities are being expanded to estimate errors in determining the missile´s position at any point along the trajectory. Inputs to BOOSTR are any combination of range, azimuth, elevation, right ascension, declination, bistatic range, or bistatic range rate from any number of space-, air-, or ground-based sensors having random errors and biases. The program also requires models of the missile, atmosphere, and gravity to compute thrust, aerodynamic drag, and mass. BOOSTR iterates using minimum variance estimation equations to obtain best estimates of missile performance parameters, including launch time, launch latitude and longitude, launch height, launch azimuth, kick angle, burnout time, and sensor data biases. Not all parameters can necessarily be estimated; those which cannot must have a priori variances given to calculate their contribution to the errors. Output from BOOSTR includes a covariance matrix of the search variables which we transform into probability error ellipsoids for the position of the missile at any time from launch to impact