Smoke Detector: Light Scattering to Detect Particle Composition and Density

Team Name: The Detectors

Members: Haosong 'Kyle' Li, Bingyu 'Teemo' Huang, Sean Brames, Shazmaan Ali, Chenghao Zhang

Team information:

Kyle started this project since this summer. We have done a small part of our design and every member learned the basic information about this project. We already have a plan and everyone knows how to approach and execute their parts .  

Mentor:

Our mentor is Henry Lee of the department of EECS. We chose Lee because of his interest and dedication to helping students and his willingness to see them succeed. We have all taken a class with Lee at some point in the past and we all like his teaching style as well as how open and welcoming he is. He is interested in the subject, and very knowledgeable in many areas of engineering, and not afraid to ask questions and open to answering them. He was our first choice as a mentor.

Project:

The end goal is to devise a system that can detect a type of gaseous particle through machine learning by detecting the scattering pattern of light when a laser is passed through the gas. We can use a laser diode to generate a light beam, and place several photodiode on the side of the beam. When there is no particle on the path of the light beam, the photodiodes will detect nothing. If there are particles in the light’s path, this will cause the light beam to scatter, and the photodetectors will detect something.We want to build a microcontroller to monitor, sample, and process the data from the hardware. Next, we can use machine learning to make it recognize different particles based on scattering patterns. I believe the team has the skills needed to accomplish the task; however, the machine learning may prove to be a challenge. The idea of this project came from EECS 188(optoelectronics) when Kyle took this course with Professor Henry Lee. This project is a good idea because it combines what we learned in classroom with our daily life. It is also very meaningful , because quick particle detection can be extremely useful in many areas. On the other hand, this is also a challenging project since we need to build the hardware and the software from scratch. We still need to work on the machine learning to train the program differentiate the different smoke such as: dust, vehicle exhaust, or smoke from cooking. To achieve this, we need to learn more about scattering and use these information to the programming and signal processing.

Work Breakdown Structure

1. Smoke detector supporting structure

   1. Build laser module fixation and aiming

   2. Build Photodiode supporting structure

   3. Build airflow system

2. Transmitter and detector circuit

   1. Build laser module control system

   2. Build photodiode-amplifier array circuits

3. Phase lock loop

   1. Design phase lock loop circuit

   2. Build output collecting system

   3. Build voltage to frequency converting circuit

   4. Build frequency to voltage converting using FPGA

4. Sampling and storing system

   1. Build sampling system

   2. Build storing system

   3. Build display system

5. Machine learning system

   1. Build machine learning processing system

6. Integration/Testing

   1. Integrate hardware and software aspect

      1. Sample data from hardware circuit and transmit the data to software part by using UART

   2. Test the whole system to see if build works.

      1. Update and return to previous steps if necessary.

7. Application

                      4.1.      Detect a type of gaseous particle when it passes through the gas

                      4.2.      Differentiate different smoke from the air, used for the study of air quality

8. Presentation

   1. Working model presentation and video

   2. Powerpoint

   3. Final Report

https://youtu.be/Yigb3vR69zA