From: Jeff Siddall <news@si...> - 2013-06-05 18:02:03
I am trying again, this time more diligently, to get occupancy
monitoring working. I have identified a few problems and an looking for
1. Hawkeyes with my W800 are plain unreliable. I would say failure to
receive sensed motion happens 25% or more of the time. This really
seems to mess up OM (garbage in, garbage out I suppose). Do other
people have problems with their W800 missing a lot of RF signals? It is
currently in the basement. Should I consider cabling the antenna higher
up in the house?
2. I am using person counts since it seems like it should be the best
option, but due to inability to steal presence from an adjoining room
(probably due to a missed motion detector) I end up with excess people.
Eventually this leads to exceeding the maximum occupancy and the
stalest one gets removed. The problem with this is it appears that
staleness is determined from the last _sensor_ report, but if the
problem is incorrect counts, then in reality the stalest person is
actually one of the incorrectly counted ones in other rooms that are
occupied by at least one "correct" person. This results in presence
being removed from a room that really is occupied (in my experience even
when motion has been sensed within the past couple of minutes!)
So how can I correct all this? I am thinking a code change to remove
stalest presence from rooms with > 1 occupant first would help. But is
using people counts hopeless if motion sensors are not reliable?