A matching algorithm for GGI aided navigation where reference map is obtained by forward method

Using forward method with the aid of DEM data is a feasible way to solve reference map problem of gradient aided navigation currently. However, there still remains a difference between the forward result and the actual gradient, which can easily lead to an estimation error of Extended Kalman Filter (EKF). The residual density is the main error resource in the forward model. After analyzes a Fourier transform forward model, we find that residual density has little influence on gradient´s fluctuation trend in local area. So a stable Nprod matching algorithm based on this characteristic is presented in the paper. The algorithm is based on the removing mean product correlation and a constraint condition is adopted at the same time. Experiment simulation at the last part of the paper approves effectiveness and validity of this method. Other characteristics can also be found out and introduced into this kind of DEM based gradient navigation system in a similar way.