Least square filtering and its application in INS/DS integrated system

To overcome the limitation of conventional Kalman filtering (CKF) in engineering application, a least square filtering (LSF) based on the state estimation is studied. LSF combines conventional least square estimation with the state estimation problem, and is without the requirement of noise statistics information. The two methods were compared using practical measuring data in laser strapdown inertial navigation system (LSINS)/DS integrated system. The simulating results shows that compared with CKF, LSF has better estimation accuracy when noise statistics information is unknown, and the convergence performance of LSF is faster, its robustness is better.