The final week of my DREU experience was spent writing up everything I had worked on over the summer! You can find it here. I am looking forward to continuing this research this semester.
This week was exciting because I was finally able to meet with an Orientation & Mobility (O&M) specialist and receive feedback on the changes I have made to AccessMap. Dr. Caspi and I met with Amy Parker, a professor at Portland State University and expert in all things orientation-mobility related.
As the program comes to an end, I’ve been polishing the features I’ve added and thinking more about how to get feedback on the application as it stands now. The three main things I worked on over the past two weeks were (1) a customizable landmark priority slider (2) more informative user interfaces for crossings, and (3) routing that takes into account the traffic control at a crossing and the presence of stairs.
I had a few breakthroughs over the past two weeks! The most exciting one was getting the routing functionality working with the modified frontend. Nick and I spent an hour trying to debug this issue a few weeks ago, but it turned out that the ordering of some of the proxy commands in a configuration file was creating the wrong proxy for certain requests. The frontend and backend are now communicating seamlessly!
Over the past two weeks, I have completed an exciting mixture of front-end and back-end tasks. I feel like I really know my way around all three repositories (
The past two weeks have been much more hands-on, as I have been getting comfortable with the AccessMap application and beginning to modify it to better fit the needs of a user with visual impairments. For the next few weeks, I will be working in a real-life “sandbox”: the Microsoft Campus in Redmond, WA. Dr. Caspi has a few contacts there who are able to add the features we are interested in to OSM. At the beginning of week 5, I used the AccessMap data pipeline to extract the pedestrian graph for the Microsoft Campus.
Over the past two weeks, my mentor and I have solidified the attributes we are interested in using for the cost function. The attributes are as follows: (1) the complexity of intersections; (2) the surface material of a pedestrian pathway; (3) the presence of obstacles on a pedestrian way, such as a fire hydrant or streetlight pole; (4) the presence of “curb bulbs”; and (5) the safety of a crossing. If I am able to successfully implement these aspects, I will shift towards thinking about the role of land use adjacent to a pedestrian way in the cost function; however, this will likely involve a more complex pipeline (accessing satellite imagery and subsequently extracting tags).
This week has been an exciting mix of hands-on work as well as some more literature review. This time, however, the literature was from a previous graduate student who was working with the deaf-blind community in Seattle in 2017-2018. I read through notes from a workshop conducted in Fall 2017 centered on spatial awareness and tactile mapping practices for deaf-blind people, as well as follow up survey responses and interviews in order to gain local contextual awareness for the project. I also read some of the articles that this graduate student had previously collected.
The first week of my DREU experience has consisted of (1) getting familiar with the existing AccessMap codebase and (2) conducting extensive research on mobile applications for visually impaired users, specifically navigation applications.