P17345: Deployable Noise Meter/public/
Customer Handoff & Final Project Documentation
Table of Contents
Team Vision for Final Demo and HandoffMSDII Final Phase Summary
- Our goals for the final phase of the project are dramatically different than the previous phases. All the testing that was able to complete, the efforts are focused on documenting the results, preparing for ImagineRIT and handing the project off. Along with presenting our project at ImagineRIT, we will also have a final review with our customer and guide and presenting a 90 second summary slide of the project to other MSD teams.
- This phase involved a lot of analysis. Presenting at ImagineRIT was successful and generated some level of interest in the project. Documenting the project's success have shown that the team mitigated some of the risks that were originally identified early in the project. However there were still many problems that were experienced. Some of these were the analog to digital converter failing to talk to the micro-controller and have an inefficient program to analyze the video of the deaf applause. It is inefficient because it takes one minute to process two seconds of the video. While we fell short on these items, we were able to prove that deaf applause is capable of being measured and paves the way for future teams to develop a more efficient program to analyze the video data.
Test Results SummaryVideo Data Summary
- It was found that a top down view of the crowd produces the best amount of data. Doing spatial frequency analysis produces FFTs that are able to analyzed and quantified. In the below figure shows an FFT with no applause.
- The figure below shows two seconds of video of people applauding. Since this was done on a Raspberry Pi, there is not enough memory to hold more video. Also since it is a Raspberry Pi the computing power is not enough to process the video fast enough to make it efficient for convocation. For 2 seconds of data it takes about 1 minute to process. However the figure below proves that it is possible to measure deaf applause.
Audio Data Summary
- For the audio portion of the project we were able to gather data from the op-amp. In the figure below shows one person in the middle of the field house with the microphone at the edge of the track. Unfortunately the analog-to-digital converter (ADC) on the board that was custom build was not able to communicate with micro-controller that was purchased. This is a real issue because if the ADC was not able to communicate with micro-controller then the data can not be used. However with the testing on the op-amp it proves that with the microphones that were purchase it is possible to measure the noise from a large distance away.
Risk and Problem Tracking
- Since our project was unable to be completed and implemented there are still a great deal of risks below shows our risks.
A live risk document can be found here: Risk Assessment
- Our risk assessment can be seen below:
- The major problem is ADC not working. The other issue is the processing time for the video. To decrease the time for the processing change out the Raspberry Pi for something else.
Live Problem Tracking Sheet: Problem Tracking
Final Project Documentation
Final Paper and Poster
- Link to Final Paper Final Paper
- Link to Poster Poster
- Link to Performance vs Requirements Performance vs Requirements
- Link to single slide Single Slide
Final Bill of Materials and Schematic
- Link to Final BOM BOM . As a note there are LEDs on the custom board that was taken from the EE senior design lab.
- Link to final Schematic Final Schematic
Plans for Wrap-up and Recommendations
- As the first phase of the project wraps up there are recommendations that the team has for the future.
- The project be re-branded. By re-branded we mean that it is geared toward NTID. There is a greater interest of quantifying the deaf applause because it has never been down (to our knowledge). Instead of the goal of the project be used at only for convocation set the project up as an access technology project. This re-branding could also gather more funding as RIT continues to develop different access technologies for the deaf/ hard- hearing culture.
- Another recommendation would be to involve the Imaging Science college. While we did talk to some professors to help us get started, a continue presence of the college (whether it be in the form of students or professors) would have helped. We were able to manage to get the initial data proving the concept, however, with input of this college the method could be improved on and perfected. Along with this do not try to process video on a Raspberry Pi. It takes to long and is inefficient
- If this project continues with the idea of it being a excitement meter for convocation then we would suggest buying different micro-controllers for the audio part. The ones that were purchased for some reason do not work with the ADC that is being used.