Hello author, Dr. @koide3,
When we creating predicted measurement sigma points from predicted measurement normal distribution N(mu, sigma), should we add measurement noise covariance into it, in code is here. But in probabilistic robotics book page 70, UKF pseudo code line 6, there is no noise covariance added, the measurement noise is added line 9. So it is a typo in code, or something i mistake? In the actual experiment, i found that the output trajectory occasionally has position jump. After repeated experimental verification for many times, i found that the phenomenon of position jump no longer appears.
Hello author, Dr. @koide3, When we creating predicted measurement sigma points from predicted measurement normal distribution N(mu, sigma), should we add measurement noise covariance into it, in code is here. But in probabilistic robotics book page 70, UKF pseudo code line 6, there is no noise covariance added, the measurement noise is added line 9. So it is a typo in code, or something i mistake? In the actual experiment, i found that the output trajectory occasionally has position jump. After repeated experimental verification for many times, i found that the phenomenon of position jump no longer appears.
