The Path Forward

We are Team AMATA, a student group from Carnegie Mellon University, MHCI program. We partnered with SWPPA and pathVu to advocate for older adults and people with disabilities, and create a pedestrian navigation experience that is safe and accessible for ALL.

It has been an honor to work with such supportive sponsors, SWPPA and pathVu, throughout this seven month project. If we can make any positive impact on accessible navigation, then we will consider this time well spent.

Currently, we have delivered a high-fidelity prototype of this accessible navigation system; but, our work doesn’t end there. We also must map out the way forward for pathVu to implement this system and successfully engage users. After all, the app cannot make a positive impact on the community if there aren’t users.

To this end, we have created a project roadmap for pathVu’s deployment of this system, outlining short, middle, and long-term stages necessary to establishing a functional service ecosystem.

Again, though, the work doesn’t stop here. There has to be a plan to expand functionality and engage more user groups — and here are some ways we’ve envisioned making this possible…

AI-Powered Photo Categorization

Currently, our prototype is designed such that data contributors first take a picture of the accessibility feature(s) they are reporting, and then they are able to select tags that apply to the picture. These tags cover general categorization of the accessibility feature (e.g. ‘ramp’, ‘stairs’) as well as more specific characteristics of these features, for instance, number of steps. Thus, the pathVu database will contain a series of images matched with specific tags which describe the contents of the images.

What does this mean? We believe the design of this reporting flow lends itself to a machine-learning approach. More specifically, a computer system would be able to create correlations between certain visual aspects of the images and the tags used to describe these images. At some point, the system will be able to automatically and accurately apply these descriptive tags. Not only does this decrease the burden on users to manually input this information, it also allows other sources of data to be leveraged and incorporated into the database. For instance, Google Street View is a large repository of visual data which includes sidewalks and storefronts. It is only a matter of time before this AI algorithm can apply descriptive tags to visual information from a service such as Street View.

In total, training this algorithm can exponentially increase the amount of data in the pathVu database, which in turn increases the value of the service to its users.

Passive Data Collection

As location tracking continues to improve, services like pathVu can continue to refine and expand their capabilities. At the moment, GPS on mobile devices is accurate enough to navigate generally, as it has a margin of error of around two to five meters. In a future where GPS precision is increased and these systems have a margin of error of a few inches, the resulting GPS data can be reliably used to make more detailed inferences. For instance, pathVu would be able to reliably detect if a user entered a building from a location other than the main entrance; this could then trigger a probe asking if the user entered an alternative entrance, and why (e.g. there was a ramp at that entrance).

Similarly, with this advanced GPS tracking, the system could also probe users about any deviations from a path, meaning users could be asked about why they avoided a certain portion of sidewalk. This advanced probing mechanism increases the ease of use for user-contributed data as well as the level of detail captured with this data. What’s more, wearables are becoming a ubiquitous vessel for digital experiences and comprise perhaps the most accessible way for users to be probed on simple questions — like, “is there a ramp here?” All of these methods increase the avenues for pathVu to collect accurate, usable data to help people navigate.

Temporary Obstacles and Conditions

One of the most difficult things to track using a service like pathVu is temporary obstacles and conditions. These range from snow and ice to construction, and can lead to significant consequences when navigating. Moving forward, integrating pathVu with other services to capture these data is a way to add a lot of value to the app. For instance, if pathVu was integrated with snowplow data, then users would be able to know which roads have been plowed, and therefore which sidewalks are most obstructed by the plowed snow. Similarly, if pathVu was integrated with weather services, then users could navigate in a way that takes weather conditions into account. For instance, a destination at the bottom of a hill after a hard rain will be less accessible to most users than a location at the top of a hill. All of this information can be taken into account to provide the user with the most comprehensive and accurate information to navigate safely.

Expanding to More Modalities

As much as we wish we could design pathVu for every single user out there, it isn’t possible on a limited timeframe. Therefore, we have focused our energy on older adults and people with mobility impairments. Thus, we’ve been working to create a functional app using a graphical user interface. Moving forward, increasing the accessibility of the app is imperative, and the logical next step is to add voice capabilities. Not only would a conversational user interface allow for users with visual impairments to use the app, it would also increase accessibility for users who want a hands free experience with the app, like people who use wheelchairs. Not all users and contexts are suited to a graphical user interface, and the app must evolve to capture these needs.

In total, there are a lot of paths forward to implement and improve this accessible navigation system. We hope our work over these past seven months have allowed pathVu to add sustained value to their app and continue to grow. Ultimately, we want to positively impact as many people as we can, empowering them to travel safely and confidently.