Close
0%
0%

Fresh Box Monitor

A self-contained monitor for cat boxes that alerts the right person to scoop or change litter

Similar projects worth following
With four cats in one house, keeping up with the litter boxes can be a challenge. This IoT device uses a microcontroller and sensors to monitor air quality and then send an alert to scoop or change the litter.

The original prototype utilized an Arduino Yun, but work is now focusing on making an ESP8266-based version which will be much more economical.

There are some possible enhancements that may come as work continues. One possibility is the use of machine learning to teach the system what air quality levels should trigger alerts. Another consideration is the use of additional sensor types to assist in identifying when to send alerts.

Litter box image credits
Green and tan box - Tom Thai, https://www.flickr.com/photos/eviltomthai/4877576787

This is the first project where I've had a 3.3V controller with a 5V sensor. I believe I have correctly used SparkFun's logic level converter, but I won't know for sure until I try it out. It's probably missing a voltage divider to get the ADC voltage down between 0 and 1V.

FreshBoxMonitor.fzz

Fritzing sketch of the wiring

fzz - 56.36 kB - 03/24/2016 at 00:43

Download

  • Prototype progress

    Mike Bohlmann04/11/2016 at 13:07 0 comments

    The ADC on the ESP8266 only converts ranges from 0 to 1V. I've decided to just power the sensor for now with 3.3V and then use a voltage divider to get a voltage range between 0 and 1V. In basic testing, this still produces usable data, but it may not be enough to get good sensitivity. I may add a second power source to see if sensitivity changes very much or use another method to get 3.3V into the ESP8266 and 5V into the sensor.

    An additional consideration is whether to use additional sensors to make it smarter. Specifically I am considering using a mixture of air quality and weight of the box. As more waste accumulates in the box, both of those factors will change. If I use both of these things with some machine learning methods, it allows the whole system to more or less skip fine tuning of sensor levels for each unit. Each unit would learn what sensor readings should trigger alerts once trained.

View project log

Enjoy this project?

Share

Discussions

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates