Information Fusion Algorithms in Ins/Smns Integrated Navigation System

As the math model and the noise statistical information are difficult to be addressed precisely and the Scene Matching Navigation System (SMNS) output is stochastic, limited and probably mismatching, only a few algorithms suitable for Inertial Navigation System /Scene Matching Navigation System (INS/SMNS) integrated navigation system for information fusion were developed. In order to find new algorithms suitable for this system, this paper studies the issue as follows. Firstly, this paper presents the improved Kalman filter by applying the methods of extrapolation, discretizing system model in unequal interval and eliminating the measurement output delay to the common Kalman filter. Secondly, this paper studies the variable step-size LMS algorithm and normalized LMS algorithm basing on  H optimal estimation. Lastly this paper applies them to the INS/SMNS integrated navigation. Simulation results demonstrate that they are suitable for INS/SMNS integrated navigation. Keywords-Information fusion, Improved Kalman filter, Variable step-size LMS, Normalized LMS