Rick

Week 1: 
Learned the context of the use case. Started working on the quick start and pretty quickly settled on a concept. Our group is now going to make a pet-like robot as a social companion. The quick start lecture was super useful to get started. I had missed the first lecture, so I missed a bit of context, but the quick start lecture got me going pretty soon.

Week 2:
Really enjoyed the guest lecture by Rebecca Schaefer! In terms of our project it was relevant to get us thinking about sound design and the importance of vocal communication/also linking this project to music. But personally it was great to get an insight into the brains of people with dementia and the progression of the illness. Rebecca seems really knowledgeable about this subject, which was great! A take-home message from this lecture that I got was: especially in late stages of dementia, it is way more important to lift up people’s spirits than to help them remember things and remind them of what they can’t do in this way.

Week 3:
Concretized our concept design a lot more. We also did a lot of literature research this week to identify what would benefit people in our use case most and to make a good research question and hypothesis for our design. We prepared the presentation for the week after, as we had to present on a Monday. I felt confident about this presentation, as we had done a lot of research into this subject and were in our opinion also delivering a realistic, usable product. We also spent quite some time identifying the relevant personas and incorporating them in our presentation. I think including personas in your design is a really useful way of thinking about users and makes you feel a little more connected to them. It may however also have some downsides, as you might be more prone to design for just that specific persona. Our product now might not work with people who are in a more severe stage of dementia or had different pets than dogs.

Week 4:
Together with Mees I did the presentation this week. It was nice doing the presentation and it felt like we could get clearly present our design. As we got our MiroCloud account this week, I started writing some simple code there together with Mees. I tried programming Miro in ROS and spent two days trying to install ROS correctly and making it work with Miro, but when that still wasn't enough to get things working, we decided to stick to MiroCloud. We then started designing our experiment in the MiroCloud simulator and writing our code in Blockly. Blockly is amazing for tiny programs and is mostly meant as an educational tool, but using it for larger programs like our experiment is a bit of a challenge and makes the code easily feel cluttered and unreadable.

Week 5:
While Mohamed and Clemente improved the Wiki, Mees and I continued working on the robot code. After a discussion with the teacher we decided to switch our use case from "going on a walk with the patient and following them" to "interacting with the patient". "following the patient" was mostly interesting from a computer science perspective and less from a social-cognitive perspective, while "interacting with the patient", really challenged us to implement social cues and design a social robot. We first had the idea to implement "petting on the head" and "petting on the back" as two functions, but noticed that Miro would sometimes register its' own movement as a "pet on the back". When we noticed this we decided to switch to "petting on the head" and "clapping" as ways to interact with Miro. Sometimes however Miro would register the sound of its' own movement as a clap. To fix this we had to turn down the reactivity of Dogg0 a lot and basically made it wait to listen for a clap a few seconds after it had stopped moving.

Week 6:
We spent a lot of time figuring out how to make Miro work on the floor. Previously we had done the experiment on a table and that worked well. When we put Miro on the floor it stopped moving forward however. It turned out that Miro saw a dark surface as a cliff and that it would not work on the black floor. Eventually we found a grey floor were one of the available Miro's would work. This week I also spent a lot of time finding a metric that reflected how much participants would trust Dogg0 and the interaction. Eventually I found an interesting paper about it and could include this in our questionnaire.

Week 7
This week we did our experiment. We found 10 participants who were willing to do the evaluation and made them fill in the questionnaire. I started analysing the results and started working on the presentation. Our experiment was not very well designed, we did not have a null-experiment and our mood score was inspired by acedemic papers, but not the same. I feel like with a better set up experiment, our project might have been a lot more rewarding. I was also hoping to have had more inspiring guest lectures like the one we got from Rebecca Schaefer by this time, I would have loved to have some more practical expert knowledge from different disciplines.

Week 8
This week we did our final presentation, which I feel we did acceptably. It was good to have a final feedback moment before finishing up the XWiki. The rest of this week we spent on improving the XWiki, polishing the details, better analysing our results and filling in missing pieces.