Integration of multi-sensor navigation data using optimal estimation techniques

Optimal filtering and smoothing algorithms are used to obtain precise aircraft position, velocity and attitude information for remote sensing applications by post-flight processing of navigation data collected by several sensors. The data from an Inertial Navigation System (INS) is differenced with data obtained from other sensors which may include photogrammetric resections, a microwave ranging system, a barometric altimeter, a radar altimeter, a laser radar, a VLF/OMEGA navigation system and doppler radar to construct error measurements. The measurements are prefiltered to compress the data and are then processed through a Kalman filter to produce estimates of the time-correlated sensor errors. The filtered error estimates are smoothed by processing backwards in time and used to correct the INS data. The track recovery program uses the UDUT covariance factorized form of the filter algorithm and is capable of processing data from any subset of sensors. The residual errors observed in processing real data collected in a number of field tests are less than 1 meter in position and less than 0.03 degrees in attitude.