The inevitability of augmented reality fitness

The more I think about it, the more I’m coming to the conclusion that fitness assistance for running and biking will be one of the killer apps for AR in the near future.

It’s hard for anyone who doesn’t run or bike semi-seriously to understand how disruptive this can be. It wasn’t for me until I started running.

Let’s start with the current state of fitness assistance with gadgets (mobile devices, watches, heart sensors, etc). It is, simply put, pretty poor. They’re not terrible. They do what they’re supposed to: measure distance, overall pace, calories/energy spent, climbing. But there’s a lot left to be desired (we just got used to it not being there).

For one thing, fidgeting with devices is hard. Using your phone while running is difficult as it might be stashed away securely, and dangerous if you’re on a bike. A wrist device is better, but usually provides limited information and operating it can be tricky. Both become a bit of a hurdle if you just want to focus on your run or ride.

But that’s not even the biggest problem, in my view. The biggest issue I have is really with how they gather metrics. Depending on what sensors you use, they can be pretty good about measuring things like power output, or heart rate. But, at least when running, measuring things that matter the most (speed, distance) is tricky. These devices tend to rely on GPS, which means they work on the aggregate, but not on short distances. They perform poorly in cities, around trees, under difficult atmospheric conditions, or tight corners.

This means that, when running, you can’t really see what your current pace is, just an approximated average over the past few minutes.

New capabilities, new solutions

AR technology can revolutionize this. Imagine you have camera-fed devices that use the visual feed to measure distance: like AR glasses that see what you see. It could actually tracks the ground beneath your feet, and give your an accurate reading of your pace over the last 5 seconds. It could measure your running distance more accurately.

This would not only mean accurate data: it would mean having data on your face (if you prefer), without having to rely on a device on your wrist, or periodical audio cues on your ear.

That last point matters. You see, because they tend to be inaccurate, most running apps – say, Strava – don’t really attempt to announce your pace out loud too often. They can do that on every km, or half km. Other apps can do that by time; say, every 5 minutes.

But there are lots of cases where this is not enough. When doing training runs, it’s typical to alternate short, fast runs with short, slower (recovery) runs (usually called a Fartlek run). In my case, the fast runs usually go for a minute or two, which means less than one kilometer. Controlling my pace in those runs is important, but it’s nigh impossible using Strava, and tricky using Fitbit (as it tends to average things, and is not very accurate; not to mention looking at my wrist when doing a fast run can be dangerous). Having a visible, up-to-speed, on-glasses pace would be much easier.

With AR googles this can be exploited in many different ways. Nothing says we need a display full of numbers. For running, having a pace-leader – someone running in front of you at the pace you wish to run – is perfect. It allows you to follow and adjust your speed when needed. Organized races usually have pace leaders running at different paces, to assist runners into keeping their planned pace.

Imagine, then, setting up a virtual pace leader that only you can see – a humanoid form running in 3d in front of you, and project onto the world. It would extremely helpful, and non-intrusive to boot.

The same goes for other methods of training. I could see the finish line of a predetermined sprint, or when am I supposed to switch my pace.

World-projected visual cues can also help with seamless routing. If you don’t have a predefined route on a known location for a long run or a bike ride, you’ll find yourself constantly fidgeting with devices to decide where to go. It’s easy to get lost, overshoot (going for a run that was longer than you planned for), or undershoot (ending up completing a workout earlier than you intended).

Imagine instead having a route on your googles, either with a pre-planned route or one that adapts with your run, giving you new time or distance estimates to select from merely by taking a right or a left. This, personally, is my dream for long distances. I like running in places I haven’t been, and pre-planning takes time (and too much of it can take the fun out of it).

Of course, this opens up other unthinkable options for training. Think of running with, or against, yourself, in a previously recorded session. Or against a friend, or a well known runner or fitness trainer. It’s hard for me to see all the possibilities, but I can tell the road it opens is immense.

Sooner or later

As much as I would like to think otherwise, I wasn’t the first to think of this. It’s interesting how many others suggested similar ideas.

There’s, of course, Ghost Pacers, a Kickstarter project that promised virtual pace leaders. I am personally extremely skeptical on their ability to deliver on their promises, but I’ll wait and see.

There’s also Snap’s spectacles. One of their developers posted an interesting video on Twitter recently, showing how the technology can work (and is supposedly already working) using their headset SDK.

The fact that similar ideas have been popping up point me to the obviousness of it. People in the intersection of technology and fitness want it, and are pushing for it.

There are many challenges that lie ahead of this dream. We don’t have commercial AR devices that sit on your face, at least not light enough ones. We’re likely years from them. Devices meant to be used during a run or a bike ride have challenges of their own, since they need to be light, impervious to the elements (or your own sweat), and have good battery life, specially if we expect them to be always on. We need a software platform, an app ecosystem, and they need to play well with existing apps (think “logging to Strava or Nike Run”) and sensors. So I can’t see it yet, or where it’s going to come from.

But it feels inevitable now.