with the change to cache results inside AP::ahrs().state we no longer need to worry about the backend's attempts to project the last-known-location forwards according to amount of time elapsed since that last-known-location was calculated.