How Intel’s AI platforms are making the Olympic and Paralympic Games more accessible – NBC Los Angeles



The following content is created in partnership with Intel. It does not reflect the work or opinions of the NBC Los Angeles editorial staff. Click here to learn more about Intel.

Some 15,000 athletes and as many as 15 million spectators are expected at the Olympic and Paralympic Games this summer. That means countless things to discover among dozens of venues—and countless ways to get lost.

Wayfinding is something many of us take for granted nowadays. With the widespread availability of wayfinding applications that provide real time maps whether you’re navigating city traffic or blazing a trail through the mountains, it’s easy to forget that live maps don’t work indoors, since the relatively weak satellite signals are blocked by large structures. While that’s a problem for all users, it’s even more of a challenge for those who may be blind or vision impaired.

An indoor wayfinding solution, powered by Intel AI platforms, will be deployed at Paris 2024. But it’s only the beginning. The project has an ambitious goal of mapping all the world’s interior spaces to better serve anyone who’s ever needed to navigate an unfamiliar indoor space.

“We’re specifically looking at how Intel technology can help people with disabilities, but it’s a universal tool as well,” says Jocelyn Bourgault, Intel’s Paris 2024 Team USA and Accessibility Programs Lead. “Even people without disabilities can gain access to it and to its benefits.”

Wayfinding at the Olympic and Paralympic training sites

Two activations will be in place at the Olympic and Paralympic Games this summer, serving athletes and staff at the Team USA Training Site in Paris, and at the International Paralympic Committee headquarters in Bonn, Germany.

Within those venues, users will be able to experience rich indoor wayfinding via a smartphone app. With the ability to search locations, enter a specific destination, or explore predefined points of interest (POIs), sighted users can choose to follow a 3D camera view through the space with an overlay of arrows on a live map, while visually impaired users can choose a 2D map view, along with audio cues, or a dark view that relies entirely on audio.

Even more granular functionality can support wheelchair users who, for example, might want to choose a route that includes ramps and elevators rather than stairs, while someone on foot can select a shorter or faster path. And as with familiar mapping applications, users can set up a route that includes stops along the way, such as at a coffee kiosk on their way to a meeting.

“We’ve created paths that allow the athlete to find their way from the minute they get off the bus: let them find their coaches immediately, get into training if that’s where they’re headed or the hot tub for rehabilitation or the physical therapy room,” says Bourgault. “We’re hoping that the athletes are going to be able to use this app to really maximize their time and, you know, focus on getting that gold medal.”

Developed as part of Intel’s unified Environmental, Social and Governance efforts, the system is very different from indoor mapping systems that rely on Bluetooth beacons and has a far simpler setup. Rather than needing to place dozens of beacons in a space to provide reference points, the process requires only LiDAR scanning of a space. A walkthrough with a 360-degree Lidar Scanner is enough to capture the interior (depending on the size of the structure, that can take a team of surveyors a day).

The resulting data point cloud, capturing the geometry of the physical space, is sent to the cloud, where machine learning algorithms trained on large datasets and running on Intel® Xeon® processors translate that point cloud data into a digital twin of the space, identifying and categorizing objects and features along the way, faster and with much higher accuracy and precision than traditional methods. That data is sent back down to edge devices, where machine learning algorithms create the actual maps for users moving through the space. And it’s all powered by Intel® AI platforms from the Intel® distribution of OpenVINO™, an open source toolkit for deploying AI on systems powered by Intel® Xeon® processors.

“We’ve been working with our partners to showcase the performance of our CPUs using OpenVINO™ in their machine learning algorithms,” says Bourgault. “And so far, the test results are incredibly positive. OpenVINO™ plus Intel® Xeon® processors is a match made in heaven.”

Further scanning can add up-to-date navigation information as elements within the space change. For example, concession stands, bleachers, staging, and other temporary structures might be set up for one event and reconfigured for another. But those subsequent scans are much faster and less data intensive.

We’re specifically looking at how Intel technology can help people with disabilities, but it’s a universal tool as well. Even people without disabilities can gain access to it and to its benefits.

Jocelyn Bourgault, Intel

Finding the future of accessibility

The Olympic and Paralympic Games activations build on previous demonstrations of the technology at a university and an international airport.  And as more large-scale installations roll out, it’s clear that there are multitudes of possibilities unlocked by dependable indoor wayfinding—from transit to retail, healthcare, government, and education.

“It’s the curb-cut effect,” says Bourgault. “Think about how a curb cut helps a disabled person get from one sidewalk to another. But once they were adopted nationwide, people realized that they don’t just help people with disabilities, they help the mother with a stroller, the kid on a trike, the delivery driver with a cart of packages. And there are lots of examples where solutions that help people with a specific disability have wide-ranging benefits for the whole community.”

Future integrations will be deeper and broader. Collaborations with outdoor mapping providers will link indoor systems into the wider world of wayfinding, opening possibilities for use in transit systems, large campuses, and, of course, across all the competition and training venues of future Olympic and Paralympic Games. And they will require the sort of extremely fast, high-capacity processing capabilities such as what Intel® Xeon® processors provide.

“What we’d like to have,” says Bourgault, “is a handoff, or handshake, between this indoor navigation system and existing outdoor navigation systems. To really make the world accessible, you need collaborations to create that seamless interaction between the spaces.”

The incorporation of live data has even more potential. Sarah Vickers, head of Intel’s Olympic and Paralympic Games Program, envisions future applications that involve continual updates with live data, enabling actions such as adjusting suggested paths to account for crowded concessions or restroom lines or busy transit stops, or directing users to retail, informational, or cultural points of interest. There are countless opportunities for optimization and efficiency—and most importantly, to provide a better user experience. “There are so many different things you can do,” says Vickers. “If you can sync this with real time data. It’s going to be really helpful. The more we can feed into that to help people, the better off it’ll be.”


More from this series:

Related Posts