A new and better adaptive algorithm based on "current" statistical (CS) model

Purpose. The traditional adaptive filtering algorithm based on CS model suffers, in our opinion, from the shortcoming of its heavy dependence on the preset-value of limit acceleration which brings frequently the rapid lowering of tracking performance when the true value of limit acceleration is not equal to the value preset in CS model. The value of limit acceleration is not preset in our new algorithm; moreover, it is calculated in real time in the course of filtering. Mathematically, the traditional CS algorithm based on CS model is performed by eqs. (12) through (18) in section 1 of the full paper, but the value of limit acceleration, instead of being preset, is calculated in real time with eq. (23) in our NCS algorithm based on new CS model in section 2. Section 2 gives the concise derivation of eq. (23); essentially, the derivation uses the new information predicted for the next moment to calculate the value of limit acceleration in real time so as to improve the track performance of our NCS algorithm. The simulation results, presented in Fig. 1 through 6, and their analysis indicate that our NCS algorithm has a performance better than the traditional CS algorithm.