Directional Motion Sensing?

[ferdies]

Hi,

Experts might be able to assist. I am thinking of implementing a directional motion sensing for the system to determine if someone is exiting the door or entering the door. My initial thought is to have 2 sensors and these sensors will be triggered sequentially (A to B = exiting; B to A = entering).

Any suggestions or maybe best practice you already have implemented?

Thanks.

[Vit]

People at Smartthings/Hubitat have done it successfully and I have a setup (tubing around the sensor lenses to limit the field) that works.

The problem is that (in known to me application) the sensors are ZigBee-based. For various reasons, they seem to be faster and in general, more reliable.
I have Fibaro, Aeon6, Dome and Neo and based on their performance, I don't think they are suitable for the application (Maybe, Neo but only if no other options available)

Also, keep in mind that the sensors' recovery time will be 20-30 seconds in the best case scenario. Depending on your application requirements, this may be unacceptable.

if your entering/exiting also involves door openings, adding a door sensor would simplify things.

Of course, the above is just my 2C.

[ferdies]

Hi Vit,

I have the Nyce Curtain Motion sensor: http://nycesensors.com/products/ncz3045/

I am thinking of having one outside (Sensor A) and another one inside (Sensor B), and you're right, I can have a "door" open trigger to make it more robust. Maybe my question is how to capture the motion "timestamp" and I can compare the timing.

Thanks.

[Vit]

I do my programming outside of HS built-in Event facilities and so, capturing 'timestamps' would not be a problem.
If I were doing it inside HS, I think I could say:

IF (when Door opens) (Motion A is active) then X.

IF (when Door opens) (Motion B is active [or Motion A is NOT active]) then Y. ==> In other words, 'entering' does not necessarily require a dedicated sensor.

Also, I believe you could save the 'timestamps' as global variables but I am of no help here.

[ericg]

ferdies, you are stating a common and much debated problem. An optimum (and generally, imperfect) solution for your situation depends on many factors that you have not identified:

• How many people (and pets, if there's a pet door) might be expected to be entering or exiting?
• The sensors you would need to detect people are generally not the ones you would use to track your pets. Are you wanting to detect people entry/exit, pet entry/exit, or both?
• Is it possible that one person's exit might overlap another person's entrance? If you are throwing a large party, and you want to count how many guests have moved through your French doors to and from your deck, I give up.
• You pose the circumstances A to B, and B to A. Is it possible to have an A-A or a B-B sequence? For example, what if you are halfway out the door, and you go back to get your car keys? How would you score this scenario: A delivery person leaves a package at the door, rings the doorbell, and leaves. Hearing the doorbell, you open the door, retrieve the package, and close the door. Was that A to B followed by B to A, or nothing?
• What is the geometry of the immediate neighborhood of your door? From a sensing point of view, it would be very convenient to have a long hall on either side of the door, with the potential of placing motion detectors at either end of the hall on each side of the door. But usually, that's not an option. And even so, if multiple people are walking those halls in both directions, the problem compounds.
• What are the relevant existing sensors, and how might they be employed to help ascertain movement? For instance, the external door I use most often has a hard-wired open/closed sensor and a Z-Wave deadbolt lock. If the door is closed and the lock button on the deadbolt's keypad gets pushed, it's a pretty good bet that somebody is leaving, rather than entering. Likewise, if the same keypad registers a code that unlocks the door, you can probably assume somebody is entering.
• Sometimes you can use time as a helpful factor. Suppose you have a motion sensor that monitors most of a room with one door. If the motion sensor has said nothing for several days, and it suddenly speaks up, then it's a pretty good bet that somebody has entered through the room's door.
• If you have multiple and separate areas that are ordinarily occupied a single person, then migration from one area to the other can be inferred by motion sensors in the two areas. Your proposal is sort of a special case of this principle, but there may be ways to trip you up.
• Is the purpose of your need security, or convenience? If the latter, then you can perhaps tolerate occasional mistakes.
• What's your budget?
• If your door is located in a reasonably secure setting -- say, separating two different areas of a warehouse, then you might want to enlist RFID devices to track who is entering and leaving. Install a reader on each side of the door. An employee waves his RFID card at the first one to unlock and open the door. After passing through, he repeats the procedure at the second reader station. This scheme tells you not only which way, but who.

To summarize: The first thing is to identify your specific needs clearly. Then evaluate the relative costs and features of monitoring hardware alternatives. If you are enhancing an existing system, pay attention to the help you might get from sensors you already have. Be creative.

[ericg]

Also, keep in mind that the sensors' recovery time will be 20-30 seconds in the best case scenario. Depending on your application requirements, this may be unacceptable.

if your entering/exiting also involves door openings, adding a door sensor would simplify things.

Yes, indeed. Some of my PIR type sensors cannot report changes more often than about once a minute. You might want to consider pressure pads that you place under a rug or door mat.

[Michael McSharry]

The one that I did is described in Section 15.8 of http://mcsSprinklers.com/mcsMQTT.pdf. It is installed in the doorway between house and garage and integrated with the existing light switch so either manual or automatic On/Off is done.

There were other prior discussions on this board dating back to the HS1, HS2 and within the last year for HS3. Use of things like motion sensor are not practical because of the time it takes the motion sensor to report vs. where a person may be when moving. What is needed is something that has both the sensing and the control in the same place. I used $5 Sonoff Basic as the host as it has the inputs for switch and light beams as well as relay to switch the mains voltage for the light.

[Uncle Michael]

I am thinking of implementing a directional motion sensing for the system to determine if someone is exiting the door or entering the door.

Can you be more specific?

• What are the circumstances? (What are the places the door connects? How many people might go through the door together?)
• What do you intend to do with the determination? (Turn on a light? Arm/Disarm an alarm? Count occupants?)

[Pete]

Here traditionally have utilize wired sensors. Overdoing it a bit with PIRs and combo Microwave/PIR sensors most recently.

My preference say for the Garage / garage doors and triggered events are just utilizing door sensors to trigger events. While PIRs would you have to open a door to get in to the garage.

Also wired up some interior doors. Very discreet as the sensors are wired in to the top of the door frame and you do not see them.

Michael mentions a combo sensor and switch using MQTT which I am also testing here. I would call this a hybrid wireless / wired sensor.

The combo GDO sensor has wires going to the garage door (optionally the entrance door), wires to the pushbutton open and close, temperature sensor and can turn on and off the garage lights. You can also add a combo PIR / Microwave sensor here. It is as fast as my wired to the Alarm panel sensors that I can see with my naked eye.

Relating to use of wired sensors which are fast still had issues relating to occupation of a room sensor.

The more sensors you utilize the more granularity you can have.

[alexbk66]

BLRadar Plug-in

[Tillsy]

I do exactly that with my front door achieved with an Aeotec window/door sensor on the front door itself and the passageway having two Aeotec multi-sensors in the ceiling (one at each end). I have various events that do things such as:

• If either of the passage sensors are in a triggered state the passage lights stay on until both sensors have been in an off state for 10 seconds (also other fancy things such as I have the lights turn on "instantly" but they "fade" off)
• If the front door is opened when there is NO movement detected in the passage then that is considered an ENTRY and various things occur (including based on other conditions such as the multi-sensors showing the temp is high or low)
• If the front door is opened when there IS movement detected at (that end of) the passage then that is considered an EXIT and various other things occur (such as turning on the porch lights)
• If the front door is held open when there is or is not movement detected then various other things occur

Works fantastically for us.