Overnight Occupancy

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • Andrew Millne
    Automated Home Ninja
    • Nov 2007
    • 269

    Overnight Occupancy

    I now have an electric heater connected via a DRH module with an HVAC object controlling seperate occupied and unoccupied heating profiles. The problem is that during the night even though the room is sealed (all doors closed) the room occupancy is reverting to unoccupied because of the reset presence timer and the heating is therefore not cutting in.

    I understand the reset presence timer is designed to safeguard against false positives so how do people overcome this?
    Visit my blog and Follow me on Twitter...
  • Karam
    Automated Home Legend
    • Mar 2005
    • 863

    #2
    Are you saying that there is someone still in the room but no motion detected in reset presence duration? If so then the first suggestion would be to increase this reset time to a value which you judge is long enough to capture two successive motions. Its function is mainly to avoid rooms remaining indefinitely 'occupied' in abnormal circumstances eg. sensor gets erroneously activated in a sealed room, so a longer period than the default should generally not constitute that much of a compromise.

    The other possibility is that the room PIR sensor location is not providing sufficient coverage, in which case there may be a need at to move it or add another sensor. Its quite unusual for no motion to be detected for very long periods of time even if people are sedentary. Even if you are asleep you will very likely get the odd motion detected within every hour - as the historical data from my own bedroom sensors will testify .

    Comment

    • Gumby
      Moderator
      • May 2004
      • 437

      #3
      I have an Idratek sensor (because I think they are more sensitive) in the ceiling over any beds to try and ensure that this doesn't happen.
      ----------------------
      www.gumbrell.com

      Comment

      • chris_j_hunter
        Automated Home Legend
        • Dec 2007
        • 1713

        #4
        >I have an Idratek sensor ... in the ceiling over any beds to try and ensure that this doesn't happen ...

        interesting, 'cos that might imply doing the same over any such area, where people might linger quietly, or not-so quietly, including desks in study, sofas in lounge, the dining table, etc ... did you put it in the ceiling because that was best, or because that was most convenient ... would putting on the wall have been as good, functionally ... would you say ???

        I ask, because our strategy is different to the norm, in that just about all our electrics will not be in wall or ceiling, but run via the floor to built-in installations - eg: bed, parlour (kitchen / diner), desk, bath, etc ... so while we could have sensors above, that might not always be so easy to arrange !

        (busy plastering, even now, maybe 25% done - meaning about 1.25 tonnes of dry-powder mixed & applied ... I'm beginning, at least, to see something for my labours ... but still lots of thinking time, for details to come, including where to put such things as sensors)
        Our self-build - going further with HA...

        Comment

        • Gumby
          Moderator
          • May 2004
          • 437

          #5
          Yes - I deliberately sited them over typical lingering areas where possible, balancing that and trying to avoid them having line of sight through doorways too deep into adjacent rooms. In fact, in my study I had to add a sensor over the desk because the one in the MFP tends to get occluded by the open door.

          Based on my understanding of the beam pattern of the sensors, it seems to me a roughly central ceiling position is going to get the most efficient coverage in terms of area covered, assuming the ceilings are not very high.

          I suspect that there is a trade-off between rapidly detecting motion near doorways, and detecting motion in the wider area of the room depending on whether you live with doors closed or doors open - ie how much information Cortex is going to be given about occupancy.

          Of course, you can always have multiple sensors.
          ----------------------
          www.gumbrell.com

          Comment

          • chris_j_hunter
            Automated Home Legend
            • Dec 2007
            • 1713

            #6
            >you can always have multiple sensors ...

            thanks ... wonder if there can be too-many ?

            and how many (if any) & what type of anscilliary ones might usefully be added (in parallel, presumably) to each of the various individual Idratek modules that can take them (eg: MFPs) ?
            Our self-build - going further with HA...

            Comment

            • Karam
              Automated Home Legend
              • Mar 2005
              • 863

              #7
              It really does depend on the structure of the room and what else you have in there for other reasons. For example an MFP/DFP will typically be placed on a wall at shoulder height and in 'conventional' rooms probably often adjacent to a door way. Typically this provides good sensing of room entry but not so good for total in room coverage. In such situations an additional off the shelf PIR can be mounted for example in an appropriate ceiling corner to provide additional coverage. This can be wired to a digital input on the wall panel and also derive its power from that location so not a great deal of extra wiring or cost. So in summary good idea and not too costly to think about extra sensors whilst you can fit them easily. No doubt we will add RF sensors to our range when we can get round to this, but as one of our clients commented - with the scale of their IDRATEK installations he didn't fancy a program of replacing batteries in several tens of modules (and before you ask ... I'm not too sure that typical energy harvesting methods will be adequate for RF PIRs just yet)

              Comment

              • Karam
                Automated Home Legend
                • Mar 2005
                • 863

                #8
                Originally posted by chris_j_hunter View Post
                and how many (if any) & what type of anscilliary ones might usefully be added (in parallel, presumably) to each of the various individual Idratek modules that can take them (eg: MFPs) ?
                Best to connect any ancillary (3rd party) PIRs to individual digital inputs so that you can differentiate between them. General purpose digital inputs are available on a number of multiple function IDRATEK modules so you'll probably often find one nearby with a spare input. Alternatively you could use a more dedicated module such as the 8 input ODI.

                Comment

                • Gumby
                  Moderator
                  • May 2004
                  • 437

                  #9
                  Of course, since I have not placed sensors in other positions I don't have a point of comparison.

                  It would be interesting to round up opinions on how well sensor positioning and choice of sensor has worked for people, both functionally and any other considerations, eg pet-friendly, aesthetic etc.
                  ----------------------
                  www.gumbrell.com

                  Comment

                  • JonS
                    Automated Home Guru
                    • Dec 2007
                    • 202

                    #10
                    In a number of rooms I have MFPs near doors in typical light-switch positions but also allowed for extra sensors running a cable to the opposite corner of the room. I have been experimenting with a Pet immune PIR and whil eit ignores the cat, it is much slower to respond (esp for children) than the Idratek ones. I have used ceiling mounted Idratek sensors in teh hall etc and they are excellent but their view does tend to be obscured by ceiling mounted pendant lamp shades. IN bed rooms I think the aesthetic of sensors over the bed are dubious and would not pass WAF criteria here.
                    For areas of little movement / lingering I am planning to look at pressuremats under sofa etc. but have not done this yet.
                    Back to the orginal comments re rooms timing out - I agree that rooms timing out is a real pain, but I guess its a compromise. I think it may be worth experimenting with very long timeout >>1hr if that were possible, but 59:45 is max atm. To my mind if the sensor network is good enough for an alarm (no false positives) then it should be good enough for occupancy. But I am sure the timeout and values were introduced out of experience.
                    JonS
                    JonS

                    Comment

                    • MrFluffy
                      Automated Home Sr Member
                      • Aug 2005
                      • 79

                      #11
                      Forgive me if this isn't the case, Im not au fait with Idratek yet, but on another system we had the problem of having two occupancy sensors in a room.
                      You would typically walk past one into the coverage of the other, and after a period of time, the first sensor would report no activity and a event would be triggered, although the second sensor would still be "seeing" the movement. On occasions one became used to waving at the second sensor to re-turn on the lights as a result of this.

                      Perhaps idra has some way to reset a shared event timer on a shared set of devices to overcome this, if not, is is a potential cause?

                      Comment

                      • JonS
                        Automated Home Guru
                        • Dec 2007
                        • 202

                        #12
                        The key concept here is that Cortex is object oriented. So a room object can have as many "connections" to its occupancy status as you want without problem, and without the complexity of having to trigger events from multuple sensors.
                        I have a couple of rooms with both an idratek MPF panel based sensor plus a 3rd part one fed into a Digital input. Although the Idratek one fires it is not mapped into the room occupancy connection as I only want people not pets registered via my pet-immune PIRs.
                        Occupancy has 3 states, un occupied, maybe occupied, definately occupied.
                        There are then various timers & logic at a room level to help determine how the room occupancy changes. There is also a "delayed presence" state which can be useful as an input (connection) to another object that you don't want to trigger if occupancy is only transitory.
                        HTH
                        JonS
                        JonS

                        Comment

                        • Karam
                          Automated Home Legend
                          • Mar 2005
                          • 863

                          #13
                          Also need to understand that Cortex works out occupancy and it is this derived signal which actions functions such as switching on lights or switching heating profile. In other words it is not simply motion detection + timer based. Several signals plus knowledge of house layout and some heuristics go into the calculation of occupancy. You do have access to the basic signals however should you wish to use these for something.

                          Comment

                          • Andrew Millne
                            Automated Home Ninja
                            • Nov 2007
                            • 269

                            #14
                            Perhaps a small change to Cortex's occupancy logic would be a good idea here?

                            I have added an additional PIR optimised for detecting movement while sleeping that i tested whilst awake, the thing is i must be a really heavy sleeper because last night i went 6 hours without moving sufficiently enough to activate the PIR.

                            If using door sensors then an increased presence timer of say 8 hours would overcome this (not currently possible) but then there is the problem of false positives in an un-occupied room.

                            What i propose is for this timer to come into use then there must be movement detected within n seconds of the door closing (This is based on the assumption that people will not get straight into bed and be out for the count instantly) if there is no motion detected within n seconds then the standard presence time-out timer is used?

                            Does this seem like a reasonable solution?
                            Visit my blog and Follow me on Twitter...

                            Comment

                            • Viv
                              Automated Home Ninja
                              • Dec 2004
                              • 284

                              #15
                              The original post was regarding the loss of occupancy in relation to HVAC heating.
                              Cortex 24 HVAC object introduced a Day and Night option with regards to the presence run on period. This was primarily intended for Bedrooms.
                              The idea being that if the room becomes occupied after a specified time e.g. 10pm (Night period start) then a long run on period can be used e.g. 8 Hours. Then at say 7am (day period start) a short run on period is used and so changing to the unoccupied profile shortly after the room becomes un-occupied.

                              The room presence solution suggested is actually what is performed i.e. the 'must be there' concept. However the reset presence/occupied timer has not been extended beyond one hour.

                              Viv

                              Comment

                              Working...
                              X