The history behind WeWALK
WeWALK began as part of the Young Guru Academy (YGA), an international non-profit based in Turkey dedicated to raising young, high potential impact entrepreneurs in the field of accessibility technologies and STEM education. In 2018, WeWALK spun off from YGA and formed its own company, and in 2020, it became part of the Microsoft AI for Accessibility program.
"We've received support from almost every level of Microsoft," notes Gökhan Meriçliler, one of the co-founders behind WeWALK. "Even before we got accepted to the AI for Good cohort in the UK, we had some experience working with the Microsoft team in Turkey because they had a partnership with YGA. That's where we first met the Microsoft team, and that's where the Azure team helped us develop some of our earliest software strategy. Later, when we came over to the UK as part of the AI for Good cohort, WeWALK was already on the market, but we needed some help with long term strategy – we were looking for some guidance and experience on improving the technology. We got really excited because Microsoft was a perfect match for this."
Meriçliler continues, "Jean Marc and I spent almost 18 weeks at the Microsoft Reactor office in London, working very closely with the Microsoft for Startups team. They helped us immensely, introducing us to many different colleagues within Microsoft and helping us build our AI strategy. They've helped us mature our product and model our social impact. We even received a shoutout during a keynote speech at Future Decoded, a large Microsoft event. In an instant, everyone in the European tech community knew about us and reached out about partnership opportunities. That really put us on the map."
WeWALK also received support from Microsoft via the Microsoft One Commercial Partner Hackathon. “Being part of the Hackathon helped us develop our AI roadmap – it was an extremely fruitful experience,” notes Feghali. “It allowed us to dive deep into the workings of our proposed model, and the Microsoft team helped us rapidly acclimate to the extensive services available in Azure. Additionally, the motivation and positivity portrayed by the Hackathon organizers and participants was truly refreshing and inspirational. As a result of the Hackathon, we were able to decide on our system architecture and design our UI, marking a great step forward for our project.”
To this day, WeWALK still works closely with both YGA and Microsoft on product and strategy development. They are also open to partnering with other assistive and adaptive technology solution providers. "There's no such thing as competition in the visually impaired world," notes Feghali. "We're all making assistive tools to help people; that's the end goal. So, if we see a promising adaptive technology growing in popularity, it's in our best interest to partner with them, integrate it into WeWALK, and provide it to our customers."
Putting AI to work for WeWALK users
One of the most innovative features of the WeWALK smart cane is the way it uses AI. While in use, the cane's gyroscope and accelerometer allow it to collect mobility data on the user, tracking things like the midline of the cane, the cane swipe angle, and the number of swipes per step. "What we realized from interviewing a bunch of mobility trainers is that most of them have a lot of clients in their caseload, sometimes up to 40 or 50 people per trainer," explains Feghali. "These clients take a short mobility training course, typically about two weeks, and then go off on their own and do their own thing. Unfortunately, there's no way for the mobility trainer to follow up on this and no easy mechanism for tracking someone's progress. This is a problem though because a lot of visual impairments are progressive—some people do lose their vision entirely, it's not often the case that someone has a stable level of sight for their entire life."
Feghali continues, "So what we're doing to address this is feeding the user's mobility data back to Azure AI, comparing it with existing datasets, and beginning to identify mobility patterns. We can then hand this data back to the visually impaired person and their mobility trainer so they can track their progress."
When asked what it's been like to work with Azure AI, İpek Erdoğan, Machine Learning Engineer at WeWALK, said, "Right from the beginning, the value of Azure AI was so clear. It made it really easy for us to investigate and understand our dataset, which is one of the most important and challenging hurdles when using machine learning. We're also using the Language Understanding (LUIS) service from Microsoft to help with our natural language processing. Language processing is tough to do from scratch, so using this ready-to-go service has helped us achieve accurate results even when we give it a small amount of input data. Overall, working with Azure AI has been great."
The path ahead – WeWALK is just getting started
In just a few short years on the market, the WeWALK smart cane has improved the lives of thousands of users.
Although WeWALK would be considered a revolutionary piece of technology by any outsider, Feghali sees it as a relatively simple endeavor. "WeWALK isn't doing anything revolutionary from a technical standpoint. It's controlling an app on a smartphone, and that app is just accessing other apps via APIs. It's an exercise in simplicity. By making things accessible, like the WeWALK app and the cane's interface, we give people a little more confidence that they're going to be able to complete their journey safely and independently."
As WeWALK continues to develop its features, their next goal is getting it into the hands of the people who need it most. "We have developed partnerships with governments, charities, health-care providers, and telecom companies to offer subsidized smart cane packages below the USD600 price point," notes Meriçliler. "Recently, we signed a protocol with the Ministry of Education in Turkey aiming to deliver WeWALK to children with visual impairments at the school level. They will integrate it into their mobility training sessions, and we will reach over 3,500 students in just a few years. This will improve their confidence with the technology and will provide WeWALK with additional mobility data so we can keep improving our services."
WeWALK has a few more ideas for improvements they'd like to make, such as integration with smart city sensors. By integrating WeWALK with sensors throughout the London Tube or the New York City Subway, for example, users could more easily navigate around crowds, events, and delays. This, alongside autonomous vehicle integration, are just a few of the team's big ideas for the future.