ENHANCEMENT AND BIAS COMPENSATION IN THE EXTENDED KALMAN OBSERVER AS A PARAMETER ESTIMATOR

by MEHROTRA, SUMIT

Methods of enhancing the extended Kalman observer as a parameter estimator (EKOPE) are analyzed. An adjustment function approach is considered for online covariance adjustment scheme to improve the EKOPE parameter estimates during periods of low input activity when this input directly affects the observability of the parameter. The effects of measurement biases on the parameter estimates given by the EKOPE are also analyzed. A method using a low pass filter is suggested and demonstrated for compensation of the measurement (sensor) biases.