CS370 FA23 Term Project : Automatic Blinds

People

• Collin Conrad
• Jared Traub
• John Mulligan

This project implements a set of smart blinds that react to outdoor lighting conditions, and indoor motion detection. This way the blinds can maximize user-experience (allowing outside light in when a room is in use), and minimize hvac energy usage (by blocking light and reducing incomming heat in the summer)

Parts Used

• Raspberry Pi
• Webcam (sensor) : Used to detect motion
• Light Dependent Resistor (Sensor) : Used to detect ouside lighting conditions

More than anything, this project is a proof of concept designed to prototype this idea, and realistically, you'd want to replace some components with those more subtible for mass production and reliability. Some of these changes could include a more purpose-built actuactor, lower-power and efficient processor (something like a esp8266 or other lower power wireless-capable microcontroller would work well here), and a more power-efficient motion detection system (like a pir sensor).

That being said, this project should still have enough to prove the feasability of a device like this.

Project Architecture

The project is split into a handfull of different "segments", with each running in their own seperate process (courtesy of python's multiprocessing). Each communicates via a queue with the Main process, which acts as a message bus and the primary state machine for the entire project.

From there there are a handful of different "modules"

HAL

The hardware abstraction layer allows us to abstract any hardware details away (e.g. for simulation or porting), and write all the code in a more general way

Motion Detection

This component uses opencv to track and notify the state machine if any motion is detected on the webcam

Web interface/Logging

This compoent runs a web-server, which will allow the user to control the blinds, as well as view metrics such as how often they've been triggered, as well as potential power savings they may have gotten

Usage

1. Install all the requirments with pip (eg. pip install -r requirments.txt)

2. NOTE: for camera support also pip install opencv-python

3. cd blinds

4. python3 main.py

Because of the design, the camera and hardware support are locked behind enviroment variables, which default to mock versions if they aren't enabled.

• USECAMERA : Enable camera support (relies on opencv being installed)
• USESERIAL : Enable the serial HAL module (relies on an arduino running the provided .ino file being connected as /dev/ttyUSB0)

Sources Used

https://pimylifeup.com/raspberry-pi-light-sensor/ https://www.circuitbasics.com/pairing-a-light-dependent-resistor-ldr-with-an-arduino-uno/ https://forum.arduino.cc/t/very-slow-rolling-average/665489/7 https://askubuntu.com/questions/1408365/wsl-how-do-i-set-the-group-for-a-tty-device https://askubuntu.com/questions/112568/how-do-i-allow-a-non-default-user-to-use-serial-device-ttyusb0 https://unix.stackexchange.com/questions/14354/read-write-to-a-serial-port-without-root https://www.cyberciti.biz/faq/find-out-linux-serial-ports-with-setserial/

https://learn.microsoft.com/en-us/windows/wsl/connect-usb#install-the-usbipd-win-project

https://medium.com/@abbessafa1998/motion-detection-techniques-with-code-on-opencv-18ed2c1acfaf