Inspiration
As a kid growing up, I loved home design and house shopping. I used to design my own floorplans and visit open houses whenever I could. Unfortunately, open houses are few and far between and scheduling an appointment can often be difficult while leading to an experience marred by pushy realtors. With that in mind, we designed HousWarm with the intention of solving these fundamental issues with house shopping.
What it does
So what does HousWarm do? In essence, we wanted to capture the experience of exploring a new home without the hassle of scheduling an appointment or the pressures of viewing with a realtor. Fundamentally, what we've aimed for, is a simple and low cost robot that is capable of livestreaming video to potential buyers who may be unavailable for an in person viewing or when there isn't a realtor available. The user would be able to remotely control the robot and view the feed in their phone or computer as if they were there in person. In addition, our goal is to integrate voice assistants to allow for question answering that potential buyers may have, such as the price, size, school district, and statistics about the neighborhood. These can be manually inputted, or in the future, automatically generated using the GPS in the Dragonboard and syncing it with online data available through services such as Google Maps or Zillow.
How we built it
In order to design this project, we tackled it from two separate ends simultaneously. On one side, we had a team use the Flask framework focusing on the web development side to display information about the user's current neighborhood and nearby neighborhoods while also integrating a live video feed as well as the process of forwarding user input for robot controls.
On the other hand, we had a separate group focus on the robot design and integration. This second group also handled the code for the Dragonboard/Raspberry Pi and Arduino used to drive the robot.
Challenges we ran into
We faced a lot of issues, especially regarding the provided hardware. Our Redbot in particular was extremely hard to work with, as the hardware desk can attest to. Many of the parts are fit together through interference fits, meaning they are shoved together and held together through friction. Unfortunately, a lot of the parts did not fit well at all and the parts were often too large for the holes that they were supposed to fit into. Therefore, we actually had to borrow a file from the hardware desk and use it to assemble the remaining parts into a usable state.
In addition, we had a lot of stability issues with the dragonboard, with even the slightest displacement resulting in a loss of display output. We eventually swapped to a Raspberry Pi instead due to constant stability issues.
Accomplishments that we're proud of
Most of us had very little hardware experience coming into this project so it was great to get to work in an entirely different field from what most of us were comfortable or experienced in. As one of the few hardware teams, and considering how little hardware was available to us, we are very proud of the implementation that we were able to create in such a short time span.
What we learned
This served as a rapid introduction to hardware hacking and was a great way to experiment with the usage of the RP, Arduino, and Dragonboard devices we had in terms of generating a cohesive and working product that also drew upon our background in software.
What's next for HousWarm
Our options for HousWarm are a bit limited due to most of the limitations of this project are due to the lack of available hardware and it's hard to say if we would invest in that hardware after this Hackathon. With that said, we are interested in continuing to improve the project if time and budget allow for it in the future.
Log in or sign up for Devpost to join the conversation.