Dementia Companion

The red box represents the vicinity detection boundary, and the blue box represents the place the patient must visit.
This is a screenshot of our notification tiles.

Inspiration

Our motivation for this project was to help dementia patients. The origins of this lay in personal experience as well as the situation of aging populations in our home countries.

We started off with a discussion of an idea one of us had. The member's initial idea was to create a smart pillbox that would remind dementia patients to refill and to take medication. Then another member pitched in for us to pivot to an application that would act as a companion to the patient through reminders of daily tasks. This would include time-critical reminders such as phone bills, as well as location based reminders for tasks such as grocery shopping.

Our vision is for our application to be a prototype of a smartphone app that acts as a true companion to the user. The application is not limited to dementia patients. It is useful to anyone. However, we found that the greatest benefactors to be dementia patients.

What it does

For time critical reminders, our prototype application scrapes through the user's email to determine the nature of the task as well as its deadline. For the prototype we focused on phone bill payments, and library overdue notices. The application would use the gathered information to provide a timely reminder to the dementia patient through push notifications.

The second aspect of our application was location based reminders. The application used and Android location API for this purpose. For our prototype, we determined when the user had not shopped for a while and was close to a grocery store. The user would then be issued a suggestion to go shopping. Our application would also use the pattern of the user's movements to make intelligent guesses if the user had gone shopping. This concept could be extended to other location based reminders.

Challenges we ran into

We were originally going to utilize an Amazon Echo as part of the companion. The advantages of using the Echo would be that the dementia patient wouldn't need to use his or her phone as often and be able to speak hands free, creating an easier to use system for the patient. However, Amazon's Alexa for the Echo is not able to speak on demand by trigger with the standard API. There are other third party API's, however, they were too time consuming for us to implement for the hackathon. This caused us to pivot our idea.

Future

We were glad to come up with a reminder app that recognized what was critical and convenient for the user. We hoped it to increase its effectiveness through innovation ideas. Especially for location based reminders, the convenience of completing a task in that location would make the user more likely to follow the advice of our app.

In the future, we can see this being integrated with IoT, smart-homes, and machine learning. For example, a smart pillbox can remind the patient to take a critical medication. A smart home can keep notifying the patient that something is being cooked, and machine learning can learn what type of tasks need to be reminded of more frequently.