lighthouse crossing beam

[knmcguire]

The crossing beam method is no longer used as default as lighthouse measurement, but is used as ground truth measurement (CONFIG_DECK_LIGHTHOUSE_AS_GROUNDTRUTH). Looking at the delta value, where 0 means that the measurement of lighthouse.x/y/z is invalid, a lot of crossing beam measurements are lost as soon as the crazyflie moves further than 4 meters from one basestation, eventhough it probably still receives solid angles on the others. This if statement put's it to 0 as soon as one sensor has not received a measurement, which is something that will probably occur a lot if the crazyflie flies further out of range of one basestation. Also the averaging in that same ifstatement suggest that the position of just one sensor can already be used for position estimate.

I also discussed with @whoenig that we should remove the averaging and base this on the sensor position instead, or the standard deviation is not really accurate anymore.

Also it seems that we are not using the CONFIG_DECK_LIGHTHOUSE_AS_GROUNDTRUTH to prevent the yaw estimate of the crossing beam getting to the kalman filter, which is probably good to implement as well.

Anyway, it would be worth to fix this at one point and make the crossing beam still a reliable method for ground truth measurements!