My sister-in-law dog Shumai is undergoing acute kidney disease. Her dog needs to go to the bathroom often. As her and husband works a lot, they are late to home. Shumai went to the bathroom in the house and started peeing blood. Dog are very smart and exhibit certain behaviors like going to the door and scratching. Using products like the August door lock and door camera, we can leverage the power of AI/ML to recognize when the dog is at the door and unlatch the hatch door so that the dog can go out and do their business. I addition, I believe some low cost hardware like a load cell can be used to monitor and track the dog's weight and food ingestion.
What it does
Monitoring the dog's weight.
Monitors the dog's food consumption.
Recognizes the dog when he/she in front of the camera. Unlocks the door so that they can do their business.
Scans the Kinship Nutrional database for food that fits the Vet's calorie profile. Provides a variety of food choices as pets can be picky.
How I built it
Uses a Raspberry Pi in tandem with a load cell for monitoring the dog's weight and food consumption. Raspberry Pi is a micro Rest-API server, serving the results of the load cell to a JSON package. The JSON package is consumed by the Android handset.
Uses curl (on the laptop and will migrate to the Pi) to request August's API to turn the lock on and off as required. Future would be to use the camera's picture data and then run open CV on it.
Challenges I ran into
I typical go to the hackathon with a partner - he knows the cloud framework very well. Since I'm alone in this endeavor I ran out of resources and time quickly. Usually we split the load and it works very well. This time, I came into know I will not complete all of the project.
Accomplishments that I'm proud of
I was able to successfully connect with both the August API and Kinship API. In addition, I'm most proud of me being able to ingest the JSON package with an Android Handset - I've never done that before as I'm still learning about Android Studio and its capabilities.
What I learned
While OpenCV works very will in an ubuntu laptop, it is a huge challenge with a embedded environment like a raspberry pi. This took up a of time. Next time, I should choose a platform where I know OpenCV can easily be deployed.
I ran out of time in regards to other parts of the apps that I wanted to implement.
What's next for August Pets
Get the image from the August camera. Slim down the raspberry pi to a MCU with a wireless wifi module. Do some alpha testing with my relatives pets.