Fuzzy Adaptive Fast Desensitized Kalman Filter

The Desensitized Kalman Filter (DKF) provides a solution to handle uncertainties in dynamic systems, but its sensitivity-weighting matrix (SWM) requires careful adjustment during the filtering process. To address this, a fuzzy adaptive fast desensitized Kalman filter (FAFDKF) is proposed based on the Takagi-Sugeno model. The FAFDKF adaptively adjusts the SWM through a softening factor, which is critical for achieving both smoothness and precision in filtering process. By integrating a fuzzy logic adaptive system, the softening factor is dynamically tuned in real-time based on the predicted residual, enabling adaptive control of the fading factor. Numerical simulations demonstrate the effectiveness of the FAFDKF, showing superior estimation capability and accuracy compared to existing DKF methods.