Improve outlier detection in LPS TDOA

[krichardsson]

Today the outlier detection simply discards all measurements that seems unreasonable based on the difference in distance to anchors is grater than a threshold (300 m). We could improve this by:

1. Anchor-anchor packest that are missing will lead to a value of 0 for the rx time in the packet received by the CF. By checking for 0 in any data we use we can find and discard those.

2. Anchor-tag packets that are missing could be detected by examining the frame time. We know that the anchors are broadcasting every 2 ms and a frame time greater than 8 * 2 = 16 ms should indicate a lost packet. It is a bit crude but will probably do for now.

[krichardsson]

1. Has been implemented

2. Was a bad idea. It adds to much knowledge in the CF about the anchors and their transmit times. We think that it will be better to add a sequence number to the packets to allow the receiver to understand if there are packets missing or not.

Closing this ticket as "good enough" for now.