Analysis of Kalman filter with correlated noises under different dependence

作者：Ma Lili^*; Wang Huiran; Chen Jinguang

来源：Journal of Information and Computational Science, 2010, 7(5): 1147-1154.

• Correlated noise
• Correlation coefficient
• Discrete linear systems
• Filtering performance
• Measurement Noise
• Standard Kalman filters

In the discrete linear system, Kalman filter will be suboptimal for state estimation when they are correlated between process and measurement noise. According to the occurring time of association, they are divided into two cases: one case is the correlated noise at the same time, and the other case is the correlated noise one time apart. There are two solutions called Method A and Method B to deal with these cases. The paper analyzes the filtering performance between the methods for correlated noises and the standard Kalman filter under different dependence, and gives one experiment. From the experimental results, we know that the solutions will be better than that of standard Kalman filter, especially when the absolution of the correlation coefficient is larger. The methods will degrade to the standard Kalman filter when the correlation coefficient equals zero. These conclusions are benefit to the engineers and researches who indicate to the fields of signal processing and state estimation.

• 出版日期2010
• 单位西安电子科技大学