Unleashing the Power of Bayesian Sensors: A guide to using it in Home Assistant!
HTML-код
- Опубликовано: 10 сен 2024
- #HomeAssistant #HomeAutomation #BinaryTechLabs
The Bayesian Sensor is a great option that is underused in Home Assistant. The sensor is based on the likelihood of an event occurring based on prior knowledge of elements that may be connected with said event.
💖 ABOUT OUR CHANNEL
Welcome to Binary Tech Labs, where we provide tutorials and reviews on Home Assistant, IoT devices, Raspberry Pi and other Single Board Computers, as well as other topics.
Check out our channel here:
/ @binarytechlabs
Don’t forget to subscribe!
🔗 GitHub Links
GitHub Bayesian - github.com/Bin...
🔗 Additional Links
Google Sheet - docs.google.co...
Bayesian Sensor Integration - www.home-assis...
Bayesian Release - www.home-assis...
FOLLOW US ON SOCIAL
Get updates or reach out to Get updates on our Social Media Profiles!
Twitter: @BinaryTechLabs
Instagram: @BinaryTechLabs
i could see this being very useful when you have a large set of sensors that don't necessary relate to one another. I have on occasion tried to trigger something to happen and i couldn't figure out how to trigger it. This could help with that
Yes, it becomes very powerful with a large number of sensors
Just came across this sensor for HA... very very cool. Spreadsheet has lots of promise. Liking the AI aspect of it 🤔
Awesome, play around with it and see what you think
A very useful video, I am grateful to you for it.. I would like more videos to come
Ty, and they are
Nice topic and I believe this could be the answer to the holy grail of integrations: automatic toilet lights. Presence can be detected like this and keep the lights on via a door contact sensor and PIR sensor in the room itself, now to figure out how to turn the light off after the person exited the toilet.
Thank you for that compliment and if you are using presence detection that could turn out the lights, or you could set up a timer trigger based on the Bayesian sensor being triggered
If you don’t keep the toilet door open when you are using it, put a trigger on the door open event. If door opened and toilet light on, turn the light off. If door opened and toilet light is off, turn it on
I don't think the spreadsheet works properly. I have emulated my current bayesian sensor data into the spreadsheet and where my sensor reads a current probability of 0.3% for the room being occupied, my threshold being 0.53%, the sensor is turning on in the sheet. Based on my testing of the sensor its working fine, so i think something is off with the sheet.
I'm sorry to hear that, I haven't come across any issues with it, I took the calculations from the Bayesian data on the home assistant page, but that's not to say that I didn't introduce some human error in. I would like to fix it if I can, but I need more data from you
thank you!
You are very welcome!
Great, thanks.
No, thank you
Just as a FYI i did figure out how to target the person that started the washer or dryer and then target where that person was when the washer or dryer finished and then send the voice alert to that home assistant in the room they are in. Mostly just used a series of automations and a few helper toggles.
Nice, I'm glad to hear that, the most recent video I released, I hoped it would help you as well. What helpers did you use?
@@BinaryTechLabs just a few home assistant toggle helpers, called them the persons name and then started washer, started dryer. Like larry started washer. Then in the automations where the washier or dryer have finished i use that helper toggle to determine who started the washer or dryer then check the current location of that persons room tracker to see what room they are in and send the voice alert to that room. Works very well i used the choose automation in the actions section to pick the right home assistant and person.
Awesome!
Very helpful! Thanks!
Thank you! Means a lot to me that I can help
@@BinaryTechLabs Absolutely. Just got ESPresence working, and I think using Bayesian sensors with my PIR motion detectors will be quite helpful
😊
So you have a sensor that turns on when someone is in the room and starts the washing machine. But the entities that trigger the bayesian sensor can also be grouped to get the same result. I still dont see the reason why you need this sensor. if you make 2 binary sensors. that turn on when someone specific is in a room and one that goes on when the washer is on and you groupe them. the grouped sensor does the exact same thing but without the stupid tresholds that dont make any sense.
Hey, thanks for your feedback, and while I agree that using only 2 entities to trigger the Bayesian sensor seems silly, just keep in mind that this is only a simple example, you can add more entities and use this sensor for things like predicting when you are cooking and more complex tasks like that where groups don't make sense, plus I only demonstrated the state platform, which shares space with groups, however numeric state and template sensors do not.
This is a great video showing the basics, but I have to agree with John H. The example doesn't really illustrate the usefulness of the bayesian sensor as opposed to a simpler solution like groups or conditions. I wish the example would also illustrate that with numeric sensors or something similar. Regardless, you explained the mechanics well. Thanks for that!
Actually, using groups is not like this sensor cause, if one thing is true in the group the whole thing is true, so every time you are in your laundry room HA will think you started the dryer, my above explanation was wrong about that.
would be nice if you created time index so we dont have to watch all 14 minutes of it
Sorry about that, normally I do for my videos. I'll check if I can add them now
Thanks, great video and awesome tool, that helps alot. I've previously struggled to define a decent weights, which effectively is defining the reliability of the entity being tested. Generally presence entities have proven to be reliably unreliable. Your tool allows for much clearer visualisation of that.
Thank you for your compliments
just wondering have you tried espresense (yes with a S not a C)? its a series of esp32s that monitor room presence. Found that if you set the distance correctly and place them in the right location in a room that they can be very accurate. Don't use your phone though to slow to respond. I use bluecharm beacons for that myself.
Yes, I actually just switched over to it and have the video up on my channel, just make sure you watch the one with no music, as the music version was a bit overpowering
Prior to switching, I was using the MQTT version that ESPresense is based on
@@BinaryTechLabs sorry i was asking @gary Barclay that question, yes have seen your videos on it and that was the main reason i tried room presence detection. Was going to try room assistant which requires raspberry pi's but with the chip sortage and all it looked to be a expensive project. So esp32's it was 😀
wher a pro wow
Thank you, I think 🤔
pքɾօʍօʂʍ
I am sorry I don't know what that means :(