Skip to content

Wrong use of prediction when no measurements are available #18

@wartek69

Description

@wartek69

Hi,
Thank you for the clear code & useful repo!
I think there is a small mistake in the usage of the Kalmann filter update step.

        if not flag:  # update using prediction
            self.b = self.lastResult
        else:  # update using detection
            self.b = b
        C = np.dot(self.A, np.dot(self.P, self.A.T)) + self.R
        K = np.dot(self.P, np.dot(self.A.T, np.linalg.inv(C)))


        self.u = np.round(self.u + np.dot(K, (self.b - np.dot(self.A,
                                                              self.u))))

In the snippet above you are using a prediction as a measurement when the flag is set. I suppose this was done to keep predicting the vessel motions when a detection disappears for a short amount of time. I do not think it is correct since the prediction will then be overconfident as it will be using the measurement noise instead of the process noise.

The correct way of handling a situation when no detections are available for some frames is to use the predict step instead of the update step.

Is my understanding correct or am I missing something?
The link to the code above is here:
https://github.com/srianant/kalman_filter_multi_object_tracking/blob/625e6c0965bb777a888ab46ed70978ad9fc93774/kalman_filter.py#L68C1-L103C22

Br
Alex

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions