Categories
REU Experiences

A nifty nine

This is my last week at Nimbus and I’m pretty bummed I won’t be working with a lot of the people I met here again. We gave practice presentations earlier this week and I was pretty impressed by all the work the other REU students got done and how nice their presentations are. The people who worked in the field, Amy and Jason, got some pretty nice work done and their presentations were super cool to hear about, especially since I’ve only been chatting with them about their work casually all summer. It’s pretty exciting to be going home so soon, too; I miss the mountains, blue lakes, and Oregon. But, I know when I get back home I’ll miss the students in my lab as well as the friends I made here. Tonight, we’re going on a bike ride to this forested area about 30 or 45 minutes away. It’s the same one from week two or three and the few of us who went then want to share the pretty bridge and canopied trails with the students who didn’t make it last time. We’re all trying to shove as much social time as possible into these last four days in Nebraska.

I got to learn so much this summer. Foremost, I learned about being my own advocate as I work on a project that I don’t know much about. The hardest part for me is knowing when something is big enough or complicated enough to ask someone about, and then figuring just who I should go and ask. Beyond the different tech I got to work with (ROS, C++, etc.), I also got to become a more independent researcher, knowing which online forums I should go to searching for information about my project, from the ACM library to Google Scholar to Stack Overflow. I also got some helpful take aways for what work environments I will enjoy in the future, easing the process of narrowing down grad school in the next two years. I’m happy to have the extra bullet points on my resume but am more excited about the experience I’ve gained and the friends I’ve made.

It’ll be a bittersweet goodbye to Nebraska and NIMBUS at the going-away coffee event the professors organized for us. Lucky me; I’ll have the opportunity to stay in touch and continue writing about the research this summer after I leave Nebraska. I really look forward to seeing where this interesting research goes.

-Shannon

Categories
REU Experiences

REU week 7 and 8

The project I’ve been working on is focused on human-robot interaction. A telepresence robot approaches someone in a library and a staff member asks them a few questions. I look at aspects of their body language and the interaction itself to judge, or come up with metrics, for the quality of the interaction. Without a lot of data, only seven or eight participants whose videos I can look at, my poster is super heavy on research, and very wordy. Also, not a lot of data means no useful graphs. I put some on their; but, I have to say I feel a bit like a catfish putting a bar graph comparing the Likert values on how reliable participants find the robot based on different head positions when on of the bars is the average of just two points. But, I think I came up with a pretty nifty graphic on how different things, like torso or head position, relate to positive or negative reactions to the robot. There’s a lot of lit review happening on my poster, which is due Monday week 9. I don’t really mind, most of the time a thoroughly enjoy the writing process anyways, but I don’t think it’s going to be winning any poster competitions. I just really like papers. Mostly that only applies to art history papers (hey what’s up I’m an art major, too) because it seems most of the historians have the freedom to add a little spice to their paper, stuff like colloquialisms or snappy editorial. HRI papers (and this might apply to most CS papers since they have to be so cutting edge, they may not have to be quite as formal) have, on the whole, shared the same pizazz.

– Shannon 

Categories
REU Experiences

REU week 6

This week feels like it flew by. Starting on Monday, I felt like I had gathered a solid base to start analyzing the first set of videos my grad student collected prior to this summer. I’m a pretty detail-oriented person, so I found categorizing people’s interactions with robots really interesting. I plan on iterating over the videos a few more times, making sure I don’t skip any details and seeing if any more important categories come to mind. When it comes to Human-Robot Interaction (HRI), it’s pretty difficult to come up with a way to qualify the quality of each person’s interaction. On one level people are just hard to read in general (and I’m not socially inept, I promise); on another level, even if you ask someone upfront if they thought the experience was effective/comfortable/etc., they’re likely to exaggerate how positive their interaction was in order to … make the researcher feel better or something? But, you’d think usable data makes a researcher feel better than just being told what other people think they want to hear.

Last week, each SRP student was assigned a variety of technologies to become familiar with to then review some dummy data with as practice. I misunderstood my technology and (Trifacta) and accidentally started learning this obscure Haskell library instead. I ended up doing some work with the real Trifacta before the meeting scheduled to show my work, but not as much as i should have had. However, I had some sick motivation to learn a new language and I’ll think I’ll keep it up with Haskell. I will also be finishing up the my project with Trifacta so I can have a well rounded perspective on it.

Categories
REU Experiences

week 5

This week has been one of my most comfortable as I prepare to start analyzing data next week. I’ve been reading up on a lot of studies about human-robot interaction, looking for ways to categorize people’s responses to an approaching robot. We also have some homework for next week: a lit review, a small statistical study, a rough draft for our posters; I’m not complaining, though. I like having goals and set deadlines so I don’t have to deal with decision fatigue. I think a sense of ownership in a project, which I’m starting to feel now that I have context for what I’m working on, alleviates that listless drone when you’re working without reason. So, the two major things I’ve been doing this week, studying and working on small, set projects, are enjoyable for me. The first gives me context and a sense of purpose, the second gives me little deadlines.

Two friends and I also went on a bike ride on Wednesday, the fourth, to the “nearby” town of Elmwood, 21 miles away, for a parade. When we arrived, we found out that there was no parade in Elmwood, but there was one six miles north in Murdock. So we biked to the parade and the free watermelon served afterwards and it was such a fun time and a good workout. I don’t think I’ve ever had so much water (or seen so much corn) in a day – ever.

-Shannon

Categories
REU Experiences

Fourth Week

This week, I was excited to read some papers on human robot interaction (HRI) studies and receive some guidance from Ohmnilabs on working with their (our?) robot. My grad student working with NIMBUS’s telepresence robots sent four papers my way to read, and they were all pretty interesting. I will be analyzing videos and placing people’s reactions to a telepresence robot into categories that I develop, so mostly focused on reading for ways other researchers have qualified/quantified reactions to their robots in a non-laboratory setting. One study used a robot called a Robovie (if you haven’t seen this cutie, look it up) inside a mall to study the effectivity of an algorithm to decide who should be approached. With the work with the study I have coming up, I’m excited to do some “there’s no right answer” work. It’s hard to get to the “good-enough” point with tech, but not so much with subjective work.

This developer from Ohmnilabs also emailed me to discuss a temporary solution for controlling the Ohmni with a ROS node. Since their company will be coming out with an update with full ROS integration in August, I won’t be able to help out with a long term solution for integrating the Ohmni into NIMBUS lab’s architecture; so, my temporary solution will probably be to build a web app to use the Ohmni API, that listens to a ROS node, listening to whatever software we have to control where the Ohmni should look and go. Sound convoluted? Realistically, a developer who knows what they’re doing may take longer than three weeks to build a usable system, and that’s all I have left, and I don’t know what I’m doing. Hopefully I’ll be making more progress next week. On the other hand, I have the Fourth of July to look forward to and along with that Wednesday off a long bike ride in the country.

-Shannon

Categories
REU Experiences

The thrilling third

Week three held a multitude of learning experiences, so this week’s blog includes the following: realizing when a project is over my head, communicating what I know and don’t know to anyone who can help me, learning on the fly, and (almost completely) exhausting my knowledge and resources on a project. Since my principal investigator didn’t come home until Thursday (no shade!), I was left to do a lot of exploration on the Ohmni – this provided a huge opportunity for self-driven study and learning how to independently work my way to a possible solution. I started the week at the point where I have a bridge into the bot, but I don’t understand the system or how to communicate with the app. More or less, I spent most the week trying to figure out how to mimic packets this bot received so I could trick it into driving around, ultimately to no avail. But, I learned a whole bunch and really deepened my interest in cybersecurity, even though this is a fairly pitiful example of “hacking.” I do think the most valuable lesson to take from this week is that, beyond the frustration and lack of success and mounds of docs and forums to sift through, I really just like coding and computers. I like learning about them, so it doesn’t feel like failure when I don’t achieve a goal. “Failure” is a little embarrassing, sure, but I’ve gotten to familiarize myself with some cool tools in the process that may come in handy later – that makes the fruitless toils worth something.

The highlight of my week would definitely be this 15 or 16 mile bike ride a big group of friends and I went on the Sunday before. Occasionally, the culture of the people here feels so different from my home college. Lots of people find bar hopping fun, for one – but that’s not really my style. I prefer privacy and quiet most the time. It’s nice to find a group of people who also crave to get out of the city for a day, which is extra hard without my car. We biked down to this park area south of Lincoln and rode on some gorgeous dirt trails. It wasn’t as removed as I’d like, since my taste leans more toward the untrimmed early season hikes in Oregon that no one has cleaned up yet, but the forest itself was undeniably gorgeous. Forests in Oregon are dense once you get into a park: lots of blackberries and brush and moss. But the trails we went through had sparse trees with a translucent canopy that let in a really comfortable amount of light. The ground cover consisted mostly of this vibrantly green, big-leaf plant. It was pretty exciting to speed down the bike path crossed by knotty roots, and lovely to stop for water and gaze across the shady greenery. It was all very cozy and refreshing.

I think I’ll bring my hammock and a book next time.

 

-Shannon

Categories
REU Experiences

The Telepresence Two (Week 2)

Week two started off even more productive than the week prior; I met my grad student who I’ll be helping out this summer and my principle investigator (PI) was out for the week. So, she (my PI), gave me what was supposed to be a fairly simple job: set up a ROS node to control this telepresence robot, herein referred to as the Ohmni, on our local network. In the beginning, I was filled with hope; OhmniLabs does refer to their robot as “open.” Don’t take this to mean “open source,” though. As I came to realize, when a company sells a development kit for a robot that costs more than that robot, their probably gonna put some work into making sure you don’t develop without that kit. Since I am writing to you from the future dimension of week 4, I can tell you that working with the Ohmni is a bit out of the scope of my expertise, which may be summed up by saying “I can make a Fizz Buzz program in over two (2) languages!”

Week two did end in a bit of a success, though. I had been trying to just get into the Ohmni that week; gaining super user was trivial, so why couldn’t I SSH into it? What am I to do? No file manager and weird permissions on downloads from the internet mean I don’t know how to download anything to help me out. This thing doesn’t even have commands for viewing file properties (this irksome property will bubble up next week). Giving up on that, I moved to the Android Debugger Bridge, since the Ohmni runs Android on the Linux Kernel. Bingo! Now I can install APKs, transfer files, open a shell from my computer (buh bye to using a touch screen to type commands into the terminal emulator), etc. But that was pretty much the end of my success for the week. As much as I’d like to be a #HackerGrrrl, I’m not. So finding my way to ADB and googling around for installs and tutorials took multiple late nights.

Well, not completely the end, I also passed my flight test on these little Hubsan drones. I’m not heavy into the flying aspect of drones, but I’m told making us interns “expert” pilots is important for a grad students’ research; so, call me Samuel, because here I am.1 

Beyond work, I play volleyball with my friends. There are sand courts by the suites (and pretty much every dorm, it seems). A few of us from Nimbus also set up two intramural sand volleyball teams. Our first game is next week. It doesn’t take long to notice how little free time grad students get. They seem to work as stacks: first in, last out. Personally, I’m not sure that’s a life I’m into having, even if it’s for a short time and especially if I can’t feel so passionately about my project that it makes up for the lack of sleep and the death of my beloved hobbies. I also ride my bike a lot. The trails around Lincoln are extensive, so it’s a great way to get out of the city, get a workout in, and feel a little more at home.

 

-Shannon

 

1. For this reference see the third chapter of Samuel in The Bible. I like to make my Catholic high school teachers proud.

Categories
REU Experiences

My first week at Nimbus

Friday, June 8th: I have big news. I found the most comfortable way to sit in Nimbus Lab, which is important when you have more than a few ROS1 tutorials to get through. The real trick is to claim a chair early to put your feet up on; being proactive is important in research.

The first week in Nimbus has been productive: my2 drone flew on this fine Friday in June, overcoming a misconfig that made the thing fly like it had too many cups of coffee and not enough water. Adam, who calls all the drones birds which I find very endearing, auto-tuned my drone and it flew stably with the correct PID values. Turns out, the same problem had shown up in another undergrad’s drone and setting some variables to the magic numbers a single drone figured out while auto-tuning fixed that problem, too. Technology is cool. Assembling the drone was pretty sweet, too; soldering can be pretty relaxing, grad students were very helpful, and the directions were clear enough. The body of the ‘bird’ is DJI with a Pixhawk ‘brain.’ Really, the only frustrating part of the whole process (besides the shaking problem, fixed by Adam) was programming the Pixhawk flight controller with this software called Flight Planner. Flight Planner only runs on Windows (I <3 my Mac); so, hello, VirtualBox. Thankfully, I only had to connect and reconnect my drone to my computer – including manually selecting the usb device I want to connect to in VB – about a million times, so it wasn’t finicky at all.

Adam was also generous enough give us a control-theory-in-a-nutshell walkthrough this morning and explained proportional-integral-derivative (PID) controllers to us. I should really invest in a comp book to write down all these notes. I really wish I could learn it all faster; I can’t wait to start writing code and helping out my Grad student. They are in Costa Rica and get back on Monday. I’ve been told I’ll be helping them with telecommunication robots, which is pretty fortunate because my drone flying skills must still be at lunch or something. Also, as much as I enjoy soldering, I’m excited to write some code.

I’ve really liked the people I’ve met so far, too. Just today, a grad student got back from a conference and introduced himself to me. We chatted CS for a bit and it reminded me what an opportunity this is to figure out if I want to stand in their shoes. So far, I’m thinking maybe I do.

Signed,

Shannon Drew

  1. That’s Robot Operating System, for all those #plebs out there.
  2. The digging team’s