Narrow beam light curtain using VL53L1X TOF sensor?

[Wookbert]

Someone out here who has experience with the VL53L1X TOF sensor?

I have a motorized roller shutter gate (see attached image), which I'd like to make smart. For safety reasons I need to add a light curtain or some other sort of narrow beam sensor, so the motor stops when someone steps below/into the closing gate.

Real light curtain sensors (so IR TX rail on one side and IR RX rail on the opposite side) is a bit too expensive when one want to cover the whole height, but first and foremost not desirable in from a wiring POV.

I'd love to use a microwave radar sensor like or similar to the LD2410, but seriously doubt that the beam/detection area can be shaped to be almost »slit« shaped (I've marked the approx. detection in the image ... I really only need like roughly 20 cm in front and behind the shutter, full height of course).

Just saw the VL53L1X TOF sensor and I’m now asking myself if this might be the solution, when set to 16×4 Spads, which would roughly lead to 27° by 6.8? (In worst case I could also place the sensor centered above the gate).

Experiences? Thoughts?

[TD-er]

Hmm not sure if that will work as there seems to be some reflective surfaces and there will be obvious changes while the roller shutter is opened or closed.
So I guess you should try it.

Personally I would opt for some kind of laser/IR light beam with some reflector.
This could then also be positioned slightly at an angle with multiple reflectors so you can cover a larger height area.

[Wookbert]

Now, check this out!

People Counting Using a Single ST VL53L1X Time-of-Flight Sensor
https://www.youtube.com/watch?v=c91Ve-g0J2U

Only question is whether it is responsive enough. Still don’t quite understand how they've done it using a single sensor!? (Perhaps the two overlapping circles/cones are misleading, as I assume there’s only one cone, while the rest is done in software).

Actually this is probably more than I need, I probably only need to narrow down the Spad grid to 16x4 or so. Guess will have to give it a try.

[ThomasB]

Still don’t quite understand how they've done it using a single sensor!? (Perhaps the two overlapping circles/cones are misleading, as I assume there’s only one cone, while the rest is done in software).

The sensor can support multiple Field of Views (FOV). This advanced functionality is not available in the ESPEasy plugin. The people crossing application, through the doorway, involves two FOVs and hardware interrupts. There is a white paper at the ST web site that explains how it is done.

But you don't need to accurately count people. So your application should be much simpler and the existing plugin should be all you need.

- Thomas

[TD-er]

[...]

The sensor can support multiple Field of Views (FOV). This advanced functionality is not available in the ESPEasy plugin. The people crossing application, through the doorway, involves two FOVs and hardware interrupts. There is a white paper at the ST web site that explains how it is done.
[...]
- Thomas

Hmm you always just know how to formulate it to trigger the "hold my beer" itches, don't you?

[Wookbert]

Hmm you always just know how to formulate it to trigger the "hold my beer" itches, don't you?

Hahaha... That’s what I like about the ESPEasy project so much. Can’t wait to see it in action.

One thing though: I strongly suggest that you improve the Home Assistant integration/implementation. ESPEasy needs to be more like „one click, auto-detect“ (just like ESPHome is) instead of manual MQTT configuration, and also must be badly added to the Home Assistant Integrations Catalogue.

When it comes to ESP-based DIY projects, I’d say that most people on the Home Assistant front are using ESPHome, which in my opinion is absolutely inferior to ESPEasy, as ESPHome’s configuration is entirely text-based YAML editing, no user-friendly GUI like in ESPEasy. But when it comes to the actual integration in HA, it’s somehow the other way around.

My fear is a bit that despite the insane development efforts of a tiny team, ESPEasy doesn’t get the attention it deserves, and might dwindle away into obsolescence. Just my two cents.

[Wookbert]

There is a white paper at the ST web site that explains how it is done.

I assume this is the documentation Thomas is referring to. One needs to click the "Documentation" tab on the top!

[Ath]

One thing though: I strongly suggest that you improve the Home Assistant integration/implementation. ESPEasy needs to be more like „one click, auto-detect“ (just like ESPHome is) instead of manual MQTT configuration, and also must be badly added to the Home Assistant Integrations Catalogue.

When it comes to ESP-based DIY projects, I’d say that most people on the Home Assistant front are using ESPHome, which in my opinion is absolutely inferior to ESPEasy, as ESPHome’s configuration is entirely text-based YAML editing, no user-friendly GUI like in ESPEasy. But when it comes to the actual integration in HA, it’s somehow the other way around.

My fear is a bit that despite the insane development efforts of a tiny team, ESPEasy doesn’t get the attention it deserves, and might dwindle away into obsolescence. Just my two cents.

Have a look at this, quite complete by now, PR #5199 (Started that on December 30th last year)
If you're willing to test we can provide a test-build (name the build you need...)

[Wookbert]

Have a look at this, quite complete by now, PR #5199 (Started that on December 30th last year)
If you're willing to test we can provide a test-build (name the build you need...)

Excellent! Regarding testing: So much curiosity, so little time ... Seriously, my head and my to-do list are packed. I could try to test it with a GYBMEP280 and LD2410 in combination on an ESP32-C3, but can’t promise when.

The other part of my suggestion was to add ESPEasy to Home Assistant’s Integration Catalogue. ESPEasy is missing on that party. Here are ESPHome and Tasmota for reference:

• https://www.home-assistant.io/integrations/esphome/
• https://www.home-assistant.io/integrations/tasmota/

[Ath]

Excellent! Regarding testing: So much curiosity, so little time ... Seriously, my head and my to-do list are packed. I could try to test it with a GYBMEP280 and LD2410 in combination on an ESP32-C3, but can’t promise when.

I'll build a 4MB C3 build 'soon-ish' for you to test

The other part of my suggestion was to add ESPEasy to Home Assistant’s Integration Catalogue. ESPEasy is missing on that party. Here are ESPHome and Tasmota for reference:

We'll pursue that goal once the MQTT AutoDiscovery feature has been merged for release