My objective was to detect the person who arrived at home.
Until today, a motion sensor was used to detect an arrival, and turn on all my scripts of presence.
The main disadvantage: I can’t discriminate a person who live at home, to an ill-intentioned person.
Several options were possible:
A GPS position send by smartphone: My objective since I started in Home Assistant Universe was to be the maximum independant of the smartphone: no 4G/3G network, no battery, etc… and I haven’t always my smartphone when I go out.
A RFID badges on house keys: A good option, but when I go out with my wife, we take just 1 key, so when we come back, just one people is detected.
A face recognition: The best option to me, with the biggest advantage that people are passive. They don’t need to execute an action at arrival to be reconized. But this technology is not really mature yet, and too much environment conditions must be met to be efficient.
A biometric recognition: The second best option to me. It brings all the benefits of RFID badge, but normally, we always go out with our fingers contrary to our keys…
After reflexion (and a big fail with facial recognition), I choose:
1 fingerprint scanner
1 RFID reader (for redondancy, and for guest who come feed Sushie (our cat) in case if we are absent too much time).
To be WAF, this project required to be:
- efficient: if we need to touch 5 times the reader to detect a presence, we’ll never use it…
- fast: avoid to wait for 10 seconds before scripts turn on
- visual: see if the presence order is taken in account
- wireless communication: It must be inside the house and near the entry door
Nota: Several technologies of fingerprint exist:
- Optical fingerprint Like This
I tested it. They are pretty easy to use with a lot of documentation and examples on internet. But they take a lot a place, they are very bright, and unidirectional. If your finger is just a little turned, it won’t work.
- Capacitive fingerprint: Like This. They are less easy to use than optical sensor and no example is available on internet. But they are faster, not bright, pretty small, and omni-directional. It’s like to unlock a smartphone !
Moreover for all these models, all fingerprints are conserved in the component, nothing in the cloud !! The component reads, analyses, and compares the fingerprint to give directly the results !
x2 NodeMCU8266: like this
x1 Fingerprint Scanner: Like This
x1 RFID Reader and badges RC522: Like This
x2 Green LED: Like This
x2 Red LED (number of led depend of number of people you want to detect. You can also use a screen).
I’m actually using 2 NodeMCU8266, because not enough input/output pins are available in one card. I’m looking for how to solve this.
On the first nodemcu are connected the RFID RC522 and the 4 led: 9 pins used
On the second is connected the Fingerprint Scanner, and probably later a sensor DHT22 for hum/temp…: 2 pins used for the moment
The steps are the following:
- Record his fingerprints & Verify his fingerprints are well reconized:
- Download the following files: https://github.com/Will711990/Manage_Capacitive_Fingerprint (I put the code to play with Arduino, because it’s not available on the internet, I dealed directly with construtor to have the arduino code!)
- put the code on your 8266 equiped with fingerprint sensor [To develop]
- follow the intructions on serial monitor
- Link the results given by fingerprint scanner to Hassio with MQTT:
- Download the following file https://github.com/Will711990/Capacitive_Fingerprint_MQTT_Hassio.git
- Put the .ino code on your 8266
- Report the .yaml in your Hassio instance
The same protocol is applied with RC522.
I will complete step to step all along the time.
Don’t look at the box, I just took one received by Amazon. I’m designing one with wood !