Project EMAR

February 2, 2017

Research & Rebuild: A Transformation of EMAR

Angela Wild

Team members preparing for field tests.

The week after the field study was extremely fruitful, especially since we were able to compile our information and see exactly where EMAR’s strengths and weaknesses in a high school setting were. One thing was for certain though, there was still a long way to go.


What We Learned

From the data gathered, we were reminded that EMAR needed a lot of changes. We collected a lot from the field study and learned much about the future direction we should ideally take. A common theme amongst the feedback was the EMAR felt like it would excel at collecting information, but lacked that human element. Teenagers didn’t feel heard. They didn’t feel like they were talking to EMAR, but simply filling out a data entry form.

One proposal to resolve this was to introduce an element of variability and randomness to EMAR’s responses. Therefore, people couldn’t predict what EMAR would say, which was an issue that came up during our field study test. There were also some software bugs with EMAR, which caused sentences to cut short or some sentences fail to play noise altogether, which possibly broke user immersion. To help facilitate with the restructuring of EMAR’s script, we developed a shared EMAR spec, which could allow everyone to easily see all of EMAR’s current interactions, and also request changes be made to it.


Looking Forwards

Despite all the work still needing to be done on EMAR’s user interaction, innovation is still growing strong. The new proposed design of EMAR V3 introduces the use of Raspberry Pi and a new programming architecture. Though the dual arduino boards with the massive array of shields served their purpose, we believe the new system would help centralize the power and stability of EMAR, while also allowing for greater flexibility on future EMAR iterations.

With the parts on their way, it leads the EMAR team to an exciting prospect of a higher-fidelity design with easier editing, cleaner interaction, and greater flexibility. It’s a big step for EMAR, and hopefully everything will work together smoothly.


The double Arduino wiring of EMAR V2.

Technical Info: The new EMAR software-wise would be composed of three parts: front-end, back-end, and the system itself. The front-end would be built just like normal web-page, using HTML, CSS, Javascript, and anything else necessary. It would be the point of user interaction with EMAR. The back-end would centralize EMAR’s actions/words/etc and would communicate with the front-end to display proper responses. Finally, the system itself would merely facilitate the communication between the different parts of EMAR, such as the eyes with the Raspberry Pi. Hardware-wise, EMAR is more or less the same with the exception of the Raspberry Pi replacing the two Arduino boards and the sound shield.


What About Poor Old EMAR V2?

As it stands right now, EMAR V2 is still very functional, though with looser unsoldered connections, an unstable hardware architecture, and a buggy code base. However, EMAR still serves its purpose and already proved it was an invaluable field testing agent.

Our plan is to keep EMAR V2 around and continue gathering and collecting valuable information, since most interactions with EMAR V2 would be similar to potential future interactions with EMAR V3. So fear not, there is still quite a while left before it’s time for EMAR to head to the Ecological Momentary Assessment Robot Retirement Home.