multiTargetMixtureTracker

[RVeler]

I understood from tutorial 10 that:

I have a measurements from camera and sonar in a csv file and I want to create multiTargetMixtureTracker.
I use kalman predictor and updater and I set:

# 5. Data association components
        self.hypothesiser = PDAHypothesiser(
            predictor=self.predictor,
            updater=self.updater,
            prob_detect=0.8
        )

        self.data_associator = JPDA(self.hypothesiser)

I set 2 initiators (because I have 2 different measurement models for the camera and the sonar), I choose the appropriate initiator base on the source of the detection (if there is more smooth way to handle detections from different measurement models I would like to hear too)

# 6. INITIATOR - Creates new tracks
        self.camera_initiator = MultiMeasurementInitiator(
            prior_state=GaussianState(
                state_vector=np.array([[0], [0], [0], [0]]),
                covar=np.diag([100, 10, 100, 10]),
                timestamp=None
            ),
            measurement_model=self.camera_model,
            deleter=CovarianceBasedDeleter(covar_trace_thresh=200.0),
            data_associator=self.data_associator,
            updater=self.updater,
            min_points=3
        )

        self.sonar_initiator = MultiMeasurementInitiator(
            prior_state=GaussianState(
                state_vector=np.array([[0], [0], [0], [0]]),
                covar=np.diag([100, 10, 100, 10]),
                timestamp=None
            ),
            measurement_model=self.sonar_model,
            deleter=CovarianceBasedDeleter(covar_trace_thresh=200.0),
            data_associator=self.data_associator,
            updater=self.updater,
            min_points=3
        )

So I ran this tracker but then I got the following error:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[7], line 265
    259     plotter.fig.show()
    264 if __name__ == "__main__":
--> 265     main()

Cell In[7], line 247, in main()
    244 detections.add(det)
    246 # Process this time step
--> 247 timestamp, track = tracker.process_step_by_step(detections, timestamp)
    248 detections_set.add(det)
    249 tracks.update(track)

Cell In[7], line 184, in MultiSensorTracker.process_step_by_step(self, detections, timestamp)
    180                 unassociated_detections.remove(hyp.measurement)
    183 self._tracks -= self.deleter.delete_tracks(self.tracks)
--> 184 self._tracks |= initiator.initiate(
    185     unassociated_detections, timestamp)
    187 return timestamp, self.tracks

File c:\Users\RINAVE\AppData\Local\miniconda3\envs\rinav\Lib\site-packages\stonesoup\initiator\simple.py:217, in MultiMeasurementInitiator.initiate(self, detections, timestamp, **kwargs)
    215 for track, hypothesis in associations.items():
    216     if hypothesis:
--> 217         state_post = self.updater.update(hypothesis)
    218         track.append(state_post)
    219         associated_detections.add(hypothesis.measurement)

File c:\Users\RINAVE\AppData\Local\miniconda3\envs\rinav\Lib\site-packages\stonesoup\updater\kalman.py:293, in KalmanUpdater.update(self, hypothesis, **kwargs)
    272 r"""The Kalman update method. Given a hypothesised association between
    273 a predicted state or predicted measurement and an actual measurement,
    274 calculate the posterior state.
   (...)    290 
    291 """
    292 # Get the predicted state out of the hypothesis
--> 293 predicted_state = hypothesis.prediction
    295 # If there is no measurement prediction in the hypothesis then do the
    296 # measurement prediction (and attach it back to the hypothesis).
    297 if hypothesis.measurement_prediction is None:
    298     # Get the measurement model out of the measurement if it's there.
    299     # If not, use the one native to the updater (which might still be
    300     # none)

AttributeError: 'MultipleHypothesis' object has no attribute 'prediction'

The problem is when in the initiator.initiate there is a problem for the kalman updater to do "update" because the hypothesis is multiple and not single. So how should I use the multiTargetMixtureTracker? is it need different initiator or updater? I didn't find information regarding this in the documentation.
I would like for any help and explanations about how to use stonesoup in a smart way.
Thanks