Ceiling versus wall mount occupancy, presence, motion sensors

Hardware
1

In starting to look at the Apollo R PRO-1 and also Everything Presence Pro, I am learning that the field of view of the sensors is not a round cone. Therefore, it seems there could be a big difference between mounting either of the sensors on the ceiling or a wall.

I am looking for feedback from folks who have experimented with ceiling vs wall placement, or have experience to share. I initially had a preference for in-ceiling, but now I worry that may come with a major limitation in how much of the floor plan in each room is covered.

Other questions include experience with ceiling mounted sensors near ceiling fans, free floating beams or other ‘obstructions’, as well as vaulted or sloped ceilings.

2

I’m your guy. I feel a bit mislead by them. I bought 7. Watched the video first where he mounted a ceiling-mounted sensor on a wall. For convenience sake. But no mention of the fact that these sensors are sub-optimal on a ceiling.

I bought 7. I found that I can’t properly adjust them, the 2450 chip, using any method including their own presence card. The 2450 dot jumps all over, is inconsistent, can’t be relied on.

I would absolutely not buy this for multiple target tracking in ceiling-mounted installations. In terms of support, it was just “look at the video”. I was very disappointed with them. Their discord presence is not new-user-friendly.

I kept them because ethernet powered was important to me. I have a ceiling in a condo where access through potlights is easy, but putting 120v adapters in the ceiling is a no-no. So I wired ethernet, PoE of course, to all 7.

To be honest, I’ve disabled sensing on the 2450 and my automations really rely on the 2412 chip. Even that is not super easy to get the boundaries correct, but they are quick and do a reasonable job in detecting movement and still energy.

Fortunately I was not buying them for direction of movement, target tracking, number of person tracking etc. For those 2450 features, I would stay away. And ceiling fans would certainly be a problem.

3

I place the sensors either way, wall or ceiling depending on the effect I want. I find I have to used both mmWave radar and IR motion detection and door sensors, Using all of the above gives a far better picture then using just one.

For example, I want a sensor in the bathroom. First, I tried a radar on the ceiling above the door and found out the radar can “see” through the closed door as people walked in the hallway. Radars are like that, and if you want them not to see through walls and doors, then use an IR sensor. Use a radar for “presence” and just write the logic into your automations.

Then I found it was not perfect and would turn off the lights with a person in the bathroom shower in some odd cases. So I installed a door sensor and used the logic “If a person is in the room and the door has remained shut, then the person must still be inside”. (I do not handle the case where a person might have jumped out the bathroom window.)

The point is that to have the best result, you need redundant sensors and think about which is worse: a false negative or a false positive and design it to fail in the preferred case.

My next round of improvements will be to look at adjoining rooms. For example, no person can be in the walk-in closet without first there being a person in the bedroom, and if the closet was occupied and now is not, then there must be a person now in the bedroom. If you have enough of these checks and enough sensors, you can catch most sensor errors and even work with some dead batteries.

This is not what you asked, but the real answer is that one sensor in one perfect spot is fragile; don’t expect perfection; use redundancy. You will NEVER find the perfect sensor or the perfect location, they don’t exist.

For example, IR motion on the ceiling over the door and close to the wall, mmWave in the wall near the back, and a sensor on the door. And listen to light switches because if one moves, it proves a person is in the room.

4

To confirm, you are talking about Apollo or Everything?

I too am looking for PoE solutions. I wish we had more sensors to choose from (including different types). I don’t look forward to changing batteries, no matter how long they last.

5

That was my motivation. The only batteries I now have here are for water leak detectors and even a few of those I’ve got powered. And yes what I was talking about in the entire reply to you was about Apollo.

6

Both of those match what I measured, so this is agreement rather than a counterpoint.

The instantaneous dot does jump. Mine sits on a desk in a living room and single frames are noisy enough that I would not trust one for position either.

What changed things for me was giving up on the instantaneous dot and using the same stream as a density map instead. Four days of target_1 coordinates, 204,796 pairs, no zones drawn:

23.8% +0.03, 1.21 m my desk chair
18.0% -0.81, 3.88 m sofa, near end
10.5% -0.66, 1.81 m the other desk chair
3.2% -1.26, 4.38 m sofa, far corner

The two sofa entries are the same sofa. They stay separate down to a 3% threshold because there is a real density valley between the cushions. None of that is visible in a single frame; it only appears after a few days.

On seeing through walls, same here, and I could put a number on it. Clipping to the room bounds dropped 6.0% of points as coming from outside the room. No sensitivity setting removed those cleanly for me either; three measured distances did.

This does not help with multi-target tracking, which is what you were asking about. It is a different use of the same sensor, and if you have already run PoE to seven of them the data is sitting there.