Self-driving wheelchair, IoT tourism app tap into foreign research talent with Mitacs program
Although Nahed Belhadj is from Tunisia, she may know more about the historical landmarks and tourist attractions of Saint John, N.B. than some locals.
Belhadj, 23, spent three months in the Maritimes working on an Internet of Things (IoT) project. She built a mobile app that offers visitors information and recommendations about nearby points of interest – from museums and restaurants to retail stores – based on their real time location as they explore the city.
“Nahed’s first week here, we were already running an indoor experiment with a museum,” recalled her project advisor Dr. Monica Wachowicz, a professor of geodesy and geomatics engineering at the University of New Brunswick (UNB). “But she just jumped in right away. She came to the experiment to help out with explaining things to people.”
Belhadj arrived in Saint John in April when “it was cold and raining all the time,” she said, “but I really enjoyed it. It’s a very, very good opportunity for me.”
That opportunity came from Mitacs Globalink, a program that brings overseas students to Canada for 12-week research internships. In the past 15 years more than 10,000 internships have been completed through the program, which is a partnership between government, academic and private sector players across Canada.
As we noted in a recent story, 500 foreign students took part in Mitacs Globalink internships at 45 Canadian universities this summer. Here’s a closer look at two of those research projects.
The IoT tourism app
Belhadj developed a mobile app to interact with 40 IoT beacons set up around Saint John at points of interest like restaurants, museums, theatres, stores and playgrounds. Once a tourist uploads the app to their smartphone, the beacons detect their proximity to those points of interest and send them information about the nearest ones, such as business hours, photos and a location map.
“We know the actual current location of the tourist and we also know their real time mobility pattern of how they are discovering Saint John. So we can help them navigate and provide them with information at the right place at the right time,” said Dr. Wachowicz, who’s also the NSERC/Cisco industrial research chair and director of UNB’s People in Motion Lab.
The app can even make ‘context aware’ recommendations to users in real time. For example, if a nearby museum is already full of visitors, the app notifies the user so they can decide whether to head there immediately or wait until it’s less crowded.
Or if a user is checking out details for a park close by, the app will incorporate real time weather data. If it looks like rain is coming, the app lets the user know and suggests indoor venues to visit nearby instead.
“Our main challenge is to make the recommendations meaningful to the tourist. So we developed a way to manage their satisfaction, to analyze if they actually followed the recommendations made to them. Then we can update the recommendations in real time,” Dr. Wachowicz said.
To ensure privacy, the app platform never stores user data, and each beacon only detects a user’s proximity within a range of its signal, not their exact location.
With safeguards for user privacy and permissions in place, the app could also collect data from users such as which attractions they visited, how much time they spent there and how they got there. Dr. Wachowicz said that could possibly be combined with other real time data the city already collects – like vehicle flow and volume patterns captured by traffic cameras – “so we really create a smart city with data in real time.”
A few days ago, Belhadj returned to Tunisia, where she’ll finish her communications degree this fall.
Google News, Bing News, Yahoo News, 200+ publications
“For me, (Globalink) was a good experience,” she said. “I learned about how to develop server notifications and mobile apps that interact with the Internet of Things. I think it will help me a lot in my future.”
The self-driving wheelchair
While Belhadj’s app helps tourists discover attractions in a city that’s new to them, Xinyi Li worked on a self-driving wheelchair system that helps people move more easily through everyday life.
Li, a 23-year-old electrical engineering student from Hangzhou, China, is in Toronto from July to October for her Globalink internship. She enjoys the mix of cultures and architecture in Toronto, but has noticed some differences between research in China and here in Canada.
“In China you have to use C++ code, so I have to learn some new things here,” Li said. “In China we want to invent some innovation but here (in Canada) we try to make an application out of an existing result.”
She is creating new algorithms for an autonomous wheelchair mechanism developed by University of Toronto researchers and Orillia, Ont.’s Cyberworks Robotics Inc.
Traditional motorized wheelchairs are operated by a joystick or sip-and-puff (SNP) mechanism. But joysticks aren’t ideal for users with impaired upper body mobility, and SNP can leave users tired and winded, said Jonathan Kelly, an assistant professor at the University of Toronto Institute for Aerospace Studies.
Using sensors, artificial intelligence (AI) software and cameras that measure distances between objects, the new chair system can autonomously move, navigate and avoid obstacles.
It can also make and store a map of a space the user navigates often, like their home or office. The user then chooses specific destinations on the map like ‘kitchen’, ‘bedroom’ or ‘lobby’ and the chair automatically takes them from one to another. (To navigate to these pre-set destinations, users could select them via a touch screen, eye gaze tracking, voice commands or SNP, Kelly said.)
The chair’s AI function “is always learning about the environment and how the environment changes … It will continuously evaluate how well its map represents what the world currently looks like, and update that map,” Kelly said.
If a new couch is placed in the room, for example, the system will learn over time that it’s a permanent fixture and add that to its internal 2D map. If it detects a cat or laundry hamper, however, the chair “is intelligent enough” to recognize it as a transient obstacle and simply move around it without updating its map, Kelly said.
The system can be retrofitted to existing wheelchairs for about $1,000 to $1,500, which is much cheaper than other self-driving models. One developed at the Massachusetts Institute of Technology uses about five sensors that cost US$7,000 each.
As for the value of the Mitacs Globalink program, Kelly said it identifies badly needed foreign talent for Canada’s high tech sector, which is already struggling to fill the 216,000 new positions it will create by 2021.
“We need to recruit some of the best students from overseas and hopefully encourage them to come to Canada for graduate studies. They may decide to spend their time doing a masters or PhD here, which is one way we can drive innovation forward in Canada. To be competitive in Canada, we really need to do that.”