As of late, a desire for a modern, smarter home interconnected through the web has evolved into a high-demand feature for new homes. We decided that we would like to tap into this industry, and utilize our skills to make our own budget IoT system, HomeNet.

What it does

Our project consists of three main parts. The facial recognition and speech detection system, the intermediary server, and the device implementation library. The detector was written to be the hub of our IoT system, HomeNet. It watches for people it knows, and when it detects a registered user, listens for their requests. Upon receiving a command, it forwards the action to the intermediary server. From there, a specific registered device would be able to complete the command. The device library we wrote enables easy creation of new devices in about 10 lines of code. This lowers the barrier for new devices and brings innovation into the common Home!

How we built it

To start, we worked on the hub. That consisted of research on facial recognition and various strategies for translating speech-to-text. In developing the hub, we quickly identified different parts that we would need to implement. From there, we split up into two teams, with one team specializing on the device library and the other working on refining the GUI for the hub and developing the server.

Challenges we ran into

Both Python and macOS restricted our ability to use multithreading while utilizing input devices, therefore changes were made to use only one process. Although mainly for ease of use, the device library had to be reworked to allow users to add devices as easily as possible. Lastly, our facial detection initially was extremely lackluster, making detection close to impossible.

Accomplishments that we're proud of

The intermediary server is well optimized and can handle more than 10 concurrent devices. The final version of the device library allows for efficient development for new devices. We were able to show how our tools allow for previously unconnected devices to be integrated using additional hardware.

What we learned

We learned how to implement facial and speech recognition software, how to integrate python with OpenCv, how to work together as a team to conquer tasks throughout the project, as well as many other skills learned throughout our development.

What's next for HomeNet

HomeNet has a dream to innovate the modern home to allow everyday users to connect every device in their domain. We aspire to design wireless micro-devices that provide the users' the best at-home experience, from specific outlets to microwave ovens to even televisions!

Built With

Share this project: