Blog 5: Week of 7/5/18
Andrew Brown again! This week started in frustration, peaked with some excitement in the middle, and ended with hopeful, yet uncertain expectation. We ended off last week with the small computer having loaded all of the crazyflie software correctly, yet unable to make the drone hover because the computer hardware was out of date. With the help of Carl, I spent some time trying to find a driver that would allow the small computer to function correctly. Eventually, however, it occurred to me that we could simply upload the correct software onto the desktop connected to Vicon. Even though we would still be running through a virtual machine, this would minimize the number of times Vicon positional data had to pass in and out of a virtual machine. We believed forcing Vicon data to go in and out of more than one virtual machine was the root of the latency problems, so I was very hopeful that placing the crazyflie software directly on the Vicon virtual machine would fix the issue … but I was wrong. There was not change in the crazyflie’s behavior.
The benefit of this failure, however, was that we could check latency issues off the list. The next step was to look at the code. Dr. Elbaum and Dr. Bradley each came and talked with Riley and me for a little bit in order to point us in the right direction in terms of code analysis. Dr. Elbaum showed us the variable we probably wanted to track. He also introduced us to some terminal tools that streamline searching for specific words across many files and documents. Dr. Bradley gave us a lesson on the PID controller in order to help us better understand what the code was trying to accomplish.
Work on the crazyflie took a pause mid-week because we had a meeting where we learned that we have a lit review to write and a paper due by the end of summer. I really wish we had been told about these things in the first week, but I believe there is still plenty of time do a good job on the paper. Serios work on the crazyflie did not really resume until after the Fourth of July break on Wednesday.
Over the break I called my parents. While describing my work to them I remembered that there was a way to make sure all of the code and all of the drone’s firmware was reverted back to the exact version used by the script we have been following. I came back Thursday excited to give this a try. As I did the preliminary tests before getting the drone to hover, things seemed really promising. Inconveniences we were dealing with before I reverted the code had disappeared, and my heart beat fast as I gave the drone the hover command.
Nothing! The droned behaved exactly as it had.
I met with Dr. Bradley later that day and he gave me a clearer picture of what my research will be after I finally get the crazyflies working. He also made plans to sit down with me and help me do some trouble shooting.
Thursday ended with one more brief moment of hope. I figured out how to get the code to print the variables of the PID error function to the consul. This revealed that the author of the code input variables to the error function in the wrong order. I switched the variables, said a little prayer, and pressed hover. Nothing changed. I looked through code a little longer and then went home.
Today, Friday, I found a new scrip that has the drones to do most of the trajectory calculation as apposed to making the computer do these calculations. This script equired me to make a new virtual machine with Ubuntu 16.04, so most of the day was spent downloading the virtual machine, getting the new script to download correctly (thank you Shannon), and fighting with network connections. Dr. Bradley and I are going to meet on Monday to see if he can’t help me figure all of this out. I really hope he can.
The worst thing about this week is that I got carried away with the crazyflie stuff and forgot to submit this blog by five o’clock this afternoon. Dr. Duncan specifically told us on Wednesday that we had to have the blogs submitted by five o’clock on Friday. Hopefully I have not messed up to badly.
Socially things are still good. I would like to note that Amy is a soldering wiz!
Hopefully next week’s report will contain some good news regarding the crazyflies.