June 22, 2022

Participate in an exciting project with Microsoft!
The Microsoft – AI for Good Research Lab reached out to Professor John Raiti for an exciting project students can get involved in for the summer quarter. They are looking for 3 MSTI students to help design an open-source portable sensor that can last 14 days on a single charge and can measure environmental factors. You may also earn independent study credit for participating in this project if you wish. If interested, email SeEun Kim (kimseeun@uw.edu) by Monday, June 27 at 12pm with the following information:
 Name
 Email
 Student ID number
 Are you a master’s or PhD student?
 Whether you wish to earn academic credit or not
 1 credit of EE 599 or EE 600 depending on your student status
 Instructor will be John Raiti
See below for more information about the project: 
——————————————————————– 
 
INTRO 
The Microsoft AI for Good Research Lab is a philanthropic team of data scientists and researchers dedicated to using AI, Machine Learning and statistical modeling to tackle some of humanity’s greatest challenges. We partner with leading nonprofits, research institutions, NGOs, and governments to accelerate work across the AI for Good program portfolio—Earth, Accessibility, Humanitarian Action, Cultural Heritage, Health—as well as other pressing issues such as affordable housing, broadband access, digital skills, justice reform, legal compliance, etc. Through Microsoft AI for Health, we have an active collaboration with UW and support numerous UW School of Medicine projects and programs. 
 
VISION 
It is becoming more and more clear that a comprehensive understanding of human health must include environmental data which has typically been difficult to obtain. We need help designing an open-source portable sensor that can last 14 days on a single charge and can measure environmental factors. Ideally
to include a 3d printable case and with parts that can be assembled by high school students without soldering as a way to engage the community and be distributable to the wider world. Since we are interested in measuring the environment inside people’s homes, the device does not need to be weatherproofed nor require online logging capabilities, as long as the data could be transferred using a MicroSD card or something similar.  
 
SPECIFICATIONS 
 Overall: 
o Less than $100 per device 
o Enclosure is 3d printable and snapped together 
o Can be assembled without soldering 
o Battery life for 10 to 14 days 
o Form factor should hopefully be the size of a wallet.  
 I/O: 

o Minute resolution at least: 
 PM 1.0  
 PM 2.5  
 PM 5.0 
 PM 10 
 VOC 
 NOx 
 Temperature 
 Humidity 
 Light Spectrogram 
 Real time clock 
o MicroSD card 
o Output format in JSON or TBD 
o LCD/OLED screen to display current conditions 
 Validation 
o It would be great to have this device validated and a paper published showing
the test-retest, intradevice variation, and limits of accuracy compared to industry
standards for these measures.  

 
INITIAL THOUGHTS: 

 Sens55 seems to fit a lot of these: https://sensirion.com/products/catalog/SEN55/  
 Here is a cheap visible spectral sensor: https://www.sparkfun.com/products/14347