Maximum Coentropy Criterion based Kalman Filtering

This project is the thesis for my B.S degree in Control & Automation in Xian Jiaotong University, advised by Prof.Guanghua Zhang. The work is mainly about the improvement of Traditional Kalman Filter(KF) under Non-Gaussian noise. In the project, two kinds of non-gaussian noise are introduced as example: Mixture Gaussian Noise & Shot Noise, and Maximum Coentropy Criterion serves as the core of improved KF(MCCKF). Both mathematical derivation & coding experiments show the outstanding performance of MCCKF under given noises compared with classical KF. The Chinese version of the paper can be found here.