Posted on

Kushaan’s status report for 3/8

 

This week I worked on fine tuning the model. I found a dataset that had fire from various point of views and perspectives. My main concerns during this process were: generalization of the perspective (can the model find our use case), quantity of sata (can this actually fine tune), and generalization to picam quality.

 

I fine tuned 11n, 11s, 11m. 11n was the smallest and took half an hour to fine tune 50 epochs. The results were promising. I also took some test images of fire in pans (our use case) on my phone and it generated a clean bounding box. 11s and 11m gave pretty bad results in comparison. Much noisier bounding boxes and under detection. We decided to use 11n going forward.

 

Once the picam arrived I did some testing using that image quality. However, it performed really well and detected 95% of our testing images. I wrote a script that took images every 2 seconds and then inferred on them.

 

One limitation of the current model is it’s night time performance. For the sake of MVP, this is not a pressing concern but is something I want to focus on after MVP. This will likely involve me making custom data.

 

I also helped out on the circuit testing. On that front, we found how to get a clear wave to drive the speaker with. We saw promising results on candle testing but need the collimater to get better results.

 

Next week I plan to work on the controls loop and interfacing since we didn’t have all the necessary hardware before break. Then I will integrate this with the vision code for a rudimentary loop.

Posted on

Team Status Report for March 8th

This week we aimed to reach our MVP requirements. With this, we were able to get the speaker and amplifier components working together. The robot chassis was built, and the imaging model was able to detect different fires that we set up well. The most significant risk that could jeopardize the success is the what if case on not being able to extinguish a fire. More specifically, even with the speaker built, it was not able to extinguish the fire. However, we are currently still working towards the idea that using a collimator will be able to put it out. On the other hand, increasing the power would also be able to be tested.

There are changes to the existing design of the system that could or may not remain the same. That would be the existing RC filter that was initially implemented. We thought that this would help with getting rid of any noise and distortion, however, with testing this week, the amplifier itself was able to fine tune the signal so there was no need for the filter. Although, if needed in the future, it is still in consideration. Cost-wise, this doesn’t change in particular as all parts were able to be found in TechSpark and such.

Part A:

This product will meet the consideration of global factors as fires exist globally. No matter the country or continent, fires, especially at home and not wild, are found commonly. This product will be able to help individuals who live in an home environment extinguish fires with ease. Furthermore, it helps for those of all ages. Whether old or young, in different stages of life, tech or not technologically savvy, this device is easy to use and requires little to none involvement on the user’s end. Governments and international organizations all emphasize the importance of fire safety regulations, which is why this further fits with in consideration of global factors.

Part B:

This product does not apply with consideration of cultural factors. It does not apply as it has no part in beliefs, values, traditions, language, etc. Since it is a technical product for a technical use, it does not take part in any cultural factors. The purpose is for safety which is designed for a universal functionality which is why it considers global factors. It relies on scientific principles of fire extinguishing which makes it universal and independent of cultural factors.

Part C:

This product applies with consideration of environmental factors as one of the main reasons as to why we are using acoustic waves, is to eliminate harmful environmental consequences that traditional extinguishing methods dispose of. For example, the powder extinguishing method continues many different types of chemicals that have proven to be harmful to bodies of water, the ecosystem that takes part in the cycle of water, and the soil and ground around as well. It is hard to throw out the extinguisher as it is then considered hazardous waste, and furthermore, it would then have to be rebought to fight more fires. Our product is reusable and does not contain any harmful chemicals and does not damage any parts of the environment that it is used in.

Part A was written by Cole MacSwain. Part B was written by Kushaan Misra. Part C was written by Stephanie Mao.

Posted on

Steph’s Status Report for March 8th

This week, I helped our group try to reach our MVP requirements.

One thing was trying to pick out components for the RC filter which was initially going to be 1.2KOhm resistor and a 2.2microF since this would yield a cutoff frequency of 60.29. However, with further testing, the amplifier fine tuned the signal enough without noise and distortion which allows no need for the filter. I further investigated this and documented the total harmonic distortion graphs that I found in the amplifier chip’s data sheet in the design report. However, it may still be in consideration to be used in the future as testing continues.

Since I was not in Pittsburgh when the robot chassis came in, I was unable to work and help on it. However, I continued to help by documenting the project in the design report, putting in the use case and design requirements, implementation, and such.

This week, I plan to construct the collimator and attach to the working speaker and amplifier apparatus to see if this will put out the fire as intended. I also plan to help on test the robot and integrating the speaker apparatus with the chassis.

Posted on

Cole’s Status Report for Mar 8th.

This week I helped as our group completed a large portion of the requirements of our MVP.

I helped to successfully write code for our ESP32 microcontroller DAC to generate the 60Hz sine wave. This was not only verified on input to the Amplifier using the oscilloscope and it was verified on the output of the Amplifier to ensure that the 60Hz wave was unmodulated.

I also helped to attach and the speaker to the Amplifier and verify that we could produce a sound. The speaker was not able to put out a fire but this was to be expected as we need a column to focus the sound waves in order to extinguish a fire.

I also helped to assemble the robot chassis that finally came in. While the it has not been tested it certainly will be this coming week.

I also helped to write the Design Report we had due friday helping to organize figures and write our testing plans for various parts of our project.

This week I plan to test the robot to make sure we can accurately drive it in the directions we need. The goal by the end of the week is to make the chassis able to integrate with the Raspberry Pi.

Posted on

Steph’s Status Report February 22

This week I continued research on the circuitry part of the design which includes the ESP32, RC filter, amplifier, and the speaker. I picked up the parts that came in and the amplifier specifically needed to be more looked into. The main thing was figuring out all the inputs and outputs of the amplifier board which would take in the signal through a 3 pin connector which would in turn be the breadboard from the filter and that would in turn be connected to the ESP32. From the other end of the amplifier board, it would take in power supply which would be between 36-48V and then the right and left connectors would be connected to the speaker magnet via ground and voltage. I am looking into purchasing the malleable steel sheets soon so that we are able to build the collimator as well with the cardboard being the interior structure.

By next week, since we have the speaker magnet, I want to construct the collimator with the cardboard and the steel sheets and then have that fitted. However, since next week will be the week prior to spring break and I will be leaving relatively early in the week, I hope that we can get more testing done with the amplifier and also get the parts for the filter.

Posted on

Team Status Report for February 22

This week, the design presentation and finalization of design were completed. Apart of that, we took in feedback on the timeline of our testing and since receiving the parts, we tested the amplifier. The following images showcases our circuit. A waveform is hooked up to the breadboard and the oscilloscope is connected to the amplifier. The amplifier itself is connected to the power supply and everything is connected to a common ground.

From this test, we were able to successfully boost our output from 100mV to 600mV with a 60Hz wave. Along with testing, we took more into consideration the heating concern as well as distance between the speaker and the fire. This would need to be continuously tested in the future which is part of our Gantt schedule already.

The most significant risk currently is trying to figure out the most ideal power supply to use to power the amplifier and all the other components in our design. This is concerning since we have a maximum of 50V and 10A which is really high, however, the plan that we currently have is to have 3 LiPo batteries in series with each other that are each 14.8V which would roughly give 44.4V (as long as it is between 36-48V).

As of now, there are no changes to the existing design presented at the design presentation this week. Moving forward, the only other costs necessary are the robot kit which we are waiting a little longer to buy since we want to fully focus on reaching  MVP first.

Posted on

Cole’s Status Report for Feb 22

This week I reviewed design presentations and helped to work on ours.

After receiving parts and creating the design presentation I was able to start conceptualizing the size, material, and general design of the housing that will hold LiPo batteries and the Raspberry Pi. At the moment the box will be laser cut out of wood with the interior and exterior in aluminum foil to keep the internal temperature steady while preventing the proximity to the fire from increasing the internal temperature.

I also helped to test the amplifier using a power supply, wave generator, and an oscilloscope. Learning how to hook everything up and power it took a decent amount of time on Saturday but we ran at least 1 successful test with the amp.

Posted on

Kushaan’s Reflection for 2/22

This week, I finished setting up the PI. I was able to ssh into it and started setting up the python environment. I installed the necessary packages and troubleshooted some installations. I started running some benchmarks of the base yolo model we want to get an idea of any inference time issues. I also set up the github for the fine-tuning, proof-of-concept, and final code. I will be working on the fine tuning code on Sunday.  I set up sshing and internet on the Pi as well. Once I finish the benchmarking, I will try different serving frameworks to see if the inference time can be further reduced. While doing fine-tuning, I will be writing boilerplate code for the logging, so we can see how the ML model influences the pipeline. I also looked into manipulating GPIO on the Pi.

In addition, I worked on the circuit to test the amplifier. We wanted to make sure the amplifier worked at lower wattage, because we don’t believe we can power it at full spec. However, if it runs at lower spec, we can hit the wattage requirements of the speaker, which is the important part. I helped diagram the circuit, and use the signal generator, DC-power supply, and oscilliscope. Ultimately, we were able to boost a 60Hz wave from 100mV to 600mV. We used a very conservative amount of amps, 200mA instead of the ~2-3A we expect to be running at full power.  We need to perform further testing once we are more confident about this circuit, but it should be fine because our ESP outputs 3V, so we should hit our required ~80W into the speaker.

Posted on

Steph’s Status Report for February 15th

This week I started finalizing the design for the speaker apparatus and purchasing items. The type of amplifier was chosen (TPA3255) to be able to meet the specs of the speaker which would be 100W RMS and 8Ohm load. This chip would be able to meet the requirements with the downside of needing a higher voltage input than we anticipated so therefore we will be increasing the LiPo batteries that are needed. Another thing was adding in a simple low-pass filter before the amplifier part since we are using an ESP32-WROOM to generate the signal. I chose the ESP32 since it is able to generate the lower frequency range of 30-60Hz that we intend to test with while also outputting a sinusoidal wave.

Once I get the parts for the speaker and all the components, I want to start testing the ESP32 first and get the code working for that. I also want to get the specific parts for the filter and build that so it would be simple to connect everything further along.

Posted on

Team Status Report for Feb 15th

This week we had our first successful proof-of-concept test! Using an app on our phones we were able to generate a 60Hz wave that put out a candle several times. To ensure our test was what we believed was going on rather than simply the air from our phone speaker we played a song at full volume at the candle. What we observed was that the song at full volume was NEVER able to extinguish the fire while the 60Hz wave consistently able to.

We also worked on our design presentation this week, incorporating feedback we received from Gloria Chang to better specify the parameters and our Minimum Viable Product.

Our next steps will be presenting this week and giving feedback to other groups. We also hope to receive the parts we ordered so that we can start assembling the parts and running tests using the full size sub-woofer magnet.

 

 

A: We are aiming to build a non-invasive, reusable fire extinguisher. This addresses a significant gap in current public welfare. With fire extinguishers that exist now a human being has to identify a fire, react quickly to find the nearest fire extinguisher and then spray a foam on the fire that puts it out. Now what was previously on fire is covered in a foam spray that needs to be cleaned off. We are eliminating all of these by having a device that can identify the fire on its own, fully extinguish the fire within 20 seconds, and it uses sound waves that will not have to be cleaned after use.

B: Does not apply. Our project aims to eliminate the human factor of a single user application and thus does not have a social interaction implication.

C: Currently, fire extinguishers expire and need to be replaced regularly. We are planning to offer a solution with rechargeable batteries that will not expire. Our solution will certainly have a higher initial cost than the ~$80 that a fire extinguisher costs, but because of the longer lifespan our solution will cost less amortized over decades.