Project Progression
-
Week1
The Beginning
My official REU journey began on May 23. I meet with Dr. Dillon, as well as another student named Gustavo who began the same day. The meeting was quick and brief, he kindly gave us the code to the REU student lab and told us our project assignments. Then, I spent the rest of the day giving Gustavo a tour of the campus since he had just arrived at UF.
On Wednesday, I met with my team so we could all get acquainted with each other. On Thursday, we had our first lab meeting. The first week was slow as we waited to find out more details about our project. I really like the other students assigned to my team. I’m feeling excited to see how this summer unfolds!
“Afoot and lighthearted I take to the open road, healthy, free, the world before me.” -Walt Whitman
-
Week2
Feeling It Out
We’ve learned more about our project. At the lab meeting this week we found out that most of the other projects have PhD leads. However, we’re one of the few teams that doesn’t have a lead. This could be both positive and negative. Without close leadership it might be hard to get the project started, or even to steer it in the right direction later on. However, it also gives us more opportunities to take charge ourselves and develop better leadership skills. At the lab meeting we were also told that we could schedule one meeting with Chris Crawford, one of the PhD students most involved with the first Brain Drone Race. We spoke to him and to best suit his schedule we are going to meet with him next Monday. Both Dr. Gilbert and Chris emphasized that this meeting should be a one time “sparking” session. So we’re all hoping that the “spark” will be electrifying, if you know what I mean.
“From a little spark may burst a flame.” -Dante Alighieri
-
Week3
The spark
This week we began with high hopes and excitement. We met with Chris to finally get a run through of the equipment and tools used in the first Brain Drone Race. We received a lot of information during the 2-hour session. However, it was still not enough. We spent the rest of the week recreating the drone flights with and without the EMOTIV headset by trial and error since Chris didn’t get a chance to show us a demonstration flight. We also recorded information about the previous race and conducted an analysis of the negative and positive aspects, as well as possible improvements. We will continue familiarization with the equipment and begin our final document in the next week. The general consensus is that we’re excited to dive into this project, but there is a lot of challenges since we have limited information about the previous race.
“If you aren’t in over your head, how do you know how tall you are?” -T.S. Eliot
-
Week4
Underway
The wheels are rolling. We each wrote sections for the final document this week. My sections were “Promotion” and “Participant Recruitment”. We then met up and conducted a group review of all the sections. It went well! We are now on our way to completing the first draft of our final document. We each picked more sections to write and plan to meet up next Tuesday again to review those. So by next week the first draft should be complete. Besides that, this week we also brainstormed testing ideas. We identified three issues we want to target. The drones have a considerable amount of drift when in flight. The line follow program has low to no functionality (the drone pretty much goes in whichever direction it desires). And we aren’t sure what the optimal tolerance should be on the EmoKey. We have already picked a day to begin testing. We know it will be a long trial and error process, but it will be exciting to try out our ideas.
“An inventor fails 999 times, and if he succeeds once, he’s in. He treats his failures simply as practice shots.” -Charles F. Kettering
-
Week5
Stumbles
This week was frustrating. The group is getting tired of documentation, which is understandable. However, it means that several group members have stopped doing their part. They say they want to do more testing, but I don’t see them active with that either. It’s difficult without any formal leadership. Those that are doing their part try to remind the others to keep working and setting targets to accomplish during the week. However, at the peer level there’s only so much reminding that can be done.
As for testing, those that participated, tried flying the drone both indoors and outdoors. Some progress was made this week, but nothing significant. We are at a road block. However, next week can be a new start. I hope we are able to regroup and regain momentum. And hopefully make more progress next week both in documentation and testing. Although this week wasn’t great, I’m confident next week will be better.
“The brick walls are there for a reason. The brick walls are not there to keep us out. The brick walls are there to give us a chance to show how badly we want something.” -Randy Pausch
-
Week6
Rising
This week was significantly better from a progress standpoint. We were able to meet with Chris again and we found a crucial bug in the code! By fixing the error in the code the line follow program improved significantly. There is still a lot of room for improvement, but now we have a better idea of what to focus on in the code. We are in the process of 3D printing propeller guards for the smaller drone, the Bebop 2. The fit was measured by hand with calipers so it has been a bit of trial and error. We had originally wanted to use a 3D scanner to make this process easier, but the scanner we have access to isn’t accurate enough to help with the small measurements we need. So that is currently underway. And a member of our team was able to get the Bebop to takeoff on a keystroke, which is a great step forward!
"Just keep swimming." -Dory
-
Week7
Beautiful Bebop Work
The propeller guards for the Bebop 2 drone were finally printed at the beginning of this week...unfortunately, the design didn’t work. I had modeled them after the ones that are currently sold for the Bebop 1, however they were too heavy and the drone couldn’t even properly lift off. I redid the design and trimmed off as much weight as I could. Later in the week we were able to print this new version. We tried it and luckily it worked. The takeoff was smooth and there was no flight disruption either!
Now that the Bebop can takeoff on a keystroke, this week we also worked on getting it integrated with the Emotiv EEG headset. When we tried it out, it worked just like the DJI drone works with the headset, which is great news. We definitely ended this week on a good note with two more things working for the Bebop 2. Next week we will most likely shift our focus back to the DJI drone to make sure we are making progress on all fronts.
“You cannot swim for new horizons until you have courage to lose sight of the shore.” -William Faulkner
-
Week8
A Little Bit of Luck
After tons of requesting, we finally secured a place for testing, the University Auditorium! We tested on three different days this week. We primarily focused on the DJI Drones. We tested different hue saturation values, altitudes, and maximum correction speeds. The first day of testing was Monday. That day yielded by far the best results in terms of conclusions. However, we also tested on Wednesday and Thursday.
Since our poster presentation is next Friday, we have also begun to think about how we want to report our progress. We plan to conduct several runs with the original code and then with the current code. We will measure deviation from the line after each run, average these results and compare to see if there has been improvement. We also want to measure how many times the drone loses sight of the line completely and see if this number has reduced.
Although, it would have been nice if all three days were as successful as the first, we are really excited about the changes in behavior of the drone. It looks as though the line follower algorithm and the flight stability have really improved.
“I'm a greater believer in luck, and I find the harder I work the more I have of it.” -Thomas Jefferson
-
Week9
Poster Week With a Side of Surprise
This week was our poster presentation session. Throughout the week we finalized our poster and practiced several times. We also recorded the voiceover for the majority of the videos we plan to include in the final documentation. Now with only one week left, we want to wrap up the documentation and have the final version done before we leave.
This week we were also invited to participate in the making of a documentary featuring the work of the first Brain Drone Race Team. The documentary was being made by a french news channel. They were making a segment on technology and the future. We helped the first Brain Drone Race Team prepare for the documentary by helping them set up, helping them control the drones when the flight crew wanted to see them in action, and by demonstrating to the camera crew how the BCI part worked as well. This part of the week was interesting because it was a very unexpected surprise to our regular schedules and the camera crew was fun to work with.
“Even in the familiar there can be surprise and wonder.” -Tierney Gearon
-
Week10
The end
Our journey has ended! We spent this week wrapping up the documentation. We included many instructional videos which will hopefully make the document more user friendly. Last minute we also decided to add sections that pertained to the Bebop 2 only. We decided to do this because although the DJI Drone is more impressive and exciting to see in a race setting, the Bebop 2 is definitely much easier to use. It’s more stable in flight, travels straighter and works well indoors, which can’t really be said about the DJI Phantom 3. We want to leave it up to the users of our documentation to decide which drone they want to use, taking into account the positive and negative aspects of each.
We had our final lab meeting this Thursday and we said our goodbye’s to everyone in the lab. This summer was a great learning experience, but I’m also very excited to go home this weekend!
“Well, here at last, dear friends, on the shores of the Sea comes the end of our fellowship in Middle-earth. Go in peace!” -J.R.R. Tolkien, The Return of the King
-
Final Entry
Final Project Evaluation
Overall the project went well. At times it was difficult that it was a group of five REU students with no PhD project leader to help guide us. However, I think this challenge definitely added to the learning experience. For the drone portion of the project we concluded that the original drone used, the DJI Phantom 3, was more complicated to use and not as adept for indoor settings as the alternative drone we explored, the Bebop 2. However, we still believe that the DJI Phantom 3 has it’s merits because it is a very powerful drone and it’s extremely exciting to see in a race setting.
The results we were able to measure for the DJI drone were:
1) Improvement in flight stability following the code glitch fix.
2) Lane alignment improved by 14.5%.
3) Amount of realignment errors in 5 runs improved by 50%.
The results we were able to measure for the Bebop drone were:1) Without line follower algorithm little to no lane deviation.
2) 90% lane alignment improvement over DJI.
3) Simpler EEG headset integration than with DJI drone, because Bebop drone emits own wifi, which headset can be connected directly to.
We were also able to complete a comprehensive documentation with instructional videos that will hopefully be an asset for reproducibility purposes in future brain drone races.