Knowledge of the process noise covariance matrix is essential for the application of Kalman filtering. However, it is usually a difficult task to obtain an explicit expression of for large time varying systems. This paper looks at an adaptive