Robotics Stack Exchange is a question and answer site for professional robotic engineers, hobbyists, researchers and students. It only takes a minute to sign up.
Sign up to join this community
I'm stuck with the following problem:
In a Kalman Filter, we assume the following holds:
The state vector can be filled with values which are returned by actual sensors. The Transition matrix needs to be set up according to a motion model. Same holds for the control input matrix. But how can I determine w_t and v_t ? How do I know which values to choose for the covariance matrices Q and R ?
I hope someone can help
All noises are assumed as Gaussian models in KF and EKF. To determine the values for the measurement noise you need to collect a sample data sets from the sensors that you are using. For each data set you can calculate the variance value put it to the corresponding position of the measurement noise covariance matrix.
Usually process noise covariances are experimental values. You can use a trial and error method to find the values which reduce the error.
