Endurance sport tech: making sense of performance sensors
In the second post of this series, our Chief Technical Officer Adam Fleming chats about performance sensors in the world of cycling, swimming and running – what data you can collect, how that’s evolving and what could be next.
“Sport science” is a term that’s been around for a long time now – it usually carries connotations of white-coats and clip-boards, with high-performance athletes chasing the last 1% of their potential for place in a national team or a world record.
While there are certainly still lab-coats and incredible athletes doing things that weren’t previously thought possible, for the average weekend athlete, these “high-tech” techniques are starting to trickle-down – and a lot of them are driven by data.
Triathlon, like so many sports, is objectively measured. The degree to which you succeed or fail is determined by a simple, objective metric – how fast you swim, cycle and run the proscribed distances. You don’t get bonus points for style or artistic impression. It all comes down to the numbers on the board as you cross the line.
As I build up towards Ironman UK, I’ve become really interested (mildly obsessed) with the training data I can gather to try and make the numbers on the board as small possible. How I use that data is something I’ll be talking about in a later article – for now though, I want to focus on the wealth of data that can be collected, and on the technology that supports it.
Time, distance and speed
Across all 3 of the sports that I’m concerned with at the moment, the primary metrics are time and distance. When measured at the start and end of repeatable workouts – or races – they give you your performance benchmarks. When measured at intermediate points (or continuously) they give you the opportunity to calculate speed (or pace) and, in general terms, speed is a pretty good indicator of effort – critically important when planning and assessing training.
Measuring time is relatively simple; by the time we finish primary school (or kindergarten if you must), we understand the concept of time and the use of simple tools to measure it. Anyone who’s done any kind of athletic training quickly becomes familiar with the tyranny of the stopwatch. Since the advent of the oscillating quartz crystal at the heart of digital watches, an accurate way to measure time has been widely and cheaply available.
Distance, however, can be a little more complex.
Let’s start with swimming – indoors, it’s relatively easy to track the distance of a swim. The pool will normally (in the UK at least) be 20, 25 or 50m potential lengths, so just count your laps and do some basic arithmetic. Similarly, using the pace-clock at the end of the pool – or your trusty smart-watch, you can get a decent measurement of your speed every time you turn around at the end of the pool.
Cycling is similarly straightforward – wheels have a known circumference, and mechanisms for counting revolutions have been around for 100+ years. Multiply circumference by revolutions and you’ve got a pretty decent measure for distance. Add that to a little bit of memory and an accurate clock and it’s easy to calculate your speed in real-time.
Running and open-water swimming are somewhat more complicated. Until recently, the standard way to deal with this was to measure the distance covered using a scale map, either before or after your run/swim – then, memorise the distance at certain waypoints. This works fine if you’re only interested in working out your pace after you’ve finished – or at those waypoints if you’re willing to indulge in some arithmetic.
Fortunately, GPS has been available in smart-watches since the early 2000s. Whilst it’s not perfect (GPS signals are easily blocked by water, trees, buildings etc, so the accuracy of your positioning is variable) it’s a massive step-up over remembering that the 3rd right turn is 1.48km from the start, or that when you’re adjacent to the boat-house you’ve done about 600m. The other massive advantage is that GPS will give you a real-time position, which suddenly gives you the ability to calculate your speed or pace continuously – in training terms, that’s a game-changer.
It’s not all about speed
We said earlier that speed is a pretty good indicator of effort – and in general terms, that’s true. Moving faster generally requires more effort, but there are a variety of other factors which contribute to the effort that you’re having to exert.
When running or cycling, the gradient of the road, the wind and any other environmental factors can have a major effect on the amount of effort you need to put in to maintain the same speed. Similarly, with open-water swimming, it’s the temperature of the water and any current.
Attempting to take these other factors into account, athletes will often look to another metric which more accurately reflects the amount of effort that’s being put in – the athlete’s heart-rate, or how fast their heart is beating.
In terms of performance sensors, there are 2 main ways that heart-rate is collected:
- a chest-strap – directly measures the electrical impulses within the heart itself
- optical sensors – these shine a bright light onto a vein complex and watch the effects of the pulse on the way that light is reflected
The chest-strap is still the standard for anyone exercising, whilst optical sensors are much improved and useful when the athlete isn’t working out – as soon as they’re moving vigorously and or sweating, the sensor becomes much more unreliable.
Optical heart-rate monitors are much better for continuous monitoring; by collecting heart-rate data continuously over a longer period between actual exercise sessions, you get a far better idea of how the athlete is recovering.
Relatively recently, there’s also been research around tracking HRV – Heart Rate Variation. This looks at how steady the heart-rate is, which is believed to be a good indicator of stress.
Having a measurement of heart-rate is a pretty good indication of the effort being put out by the athlete, but sport is all about optimisation. The ideal, particularly for sports like triathlon, is to be able to move as fast as possible with as little effort as possible. One way to do this is to train the athlete to work harder with less effort – go faster with a lower heart-rate. But on its own, this can only go so far – the other axis is to look at how efficient they are at that specific sport.
Traditionally, this kind of efficiency optimisation is in the realm of sport-science research and the professional coach. We’ll talk more about that in another post, along with how software is starting to make these kinds of premium analysis available to the average athlete. But before going there, we’re talking about some of the key enabling technologies – sensors (or, for what we’re looking at, performance sensors).
Making sense of performance sensors
There are 3 main areas of innovation when it comes to performance sensors, which have had a profound effect on amateur sport:
- Sensor fusion with GPS for improved location accuracy
- Inertial sensors
- Strain gauges
Around the early 2000s, GPS-enabled watches became widely available, removing a lot of the guesswork in outdoor positioning. However, GPS (especially early GPS) isn’t completely accurate under all conditions.
A lot of innovation has gone into the modern sensor-fused positioning systems in current smart-watches. Sensor-fusion is the idea that you fuse information from one source (GPS for example) with information from another (say an accelerometer) to correct expected errors.
For example, GPS signals are subject to all kinds of interference which can result in low-accuracy positioning information. You could be running at a consistent pace, but the watch thinks you’re bounding forward in 20m jumps every 30 seconds or so.
If you’re also tracking the output of an accelerometer you can (to an extent) validate the information you’re getting from GPS to correct errors. In the case above where GPS sees you leaping forward in giant bounds, if your accelerometer information isn’t consistent with it (you don’t see huge bursts of acceleration, but rather reasonably stable movement), then you can correct your GPS signal to an extent.
This is a good point to introduce the hidden star of the show – inertial sensors.
Inertial sensors are accelerometers and gyroscopes, used to measure linear and rotary acceleration (pushes and twists if you prefer). When paired (as they often are) with magnetometers, you’ve got the bare-bones of an inertial navigation system – or INS – and an INS is what’s used in rocket-guidance systems.
As you’d imagine, the sensors in rocket-guidance packages are not typically in the price-range of the amateur athlete – but over the last few years, the technology has been miniaturised and commoditised, and can now be found in your smartphone, your smartwatch and in some performance sensors made specifically for sport.
The science (and maths) of how inertial sensors can be used to measure movement is pretty deep, and gets into some pretty nasty concepts – but for the purposes of this article, an inertial sensor package gives you the ability to monitor how a point moves through space.
So, focusing on their use as performance sensors, if you mount one in a wristwatch and swim with it, you’ll see a cyclic motion which maps to the movement of your arm. Run with the same wrist-watch and you’ll see the motion of the arm as the runner moves through the gait cycle. Put it into a foot-pod and you can see the motion of the foot and thus measure things like how long the foot is in contact with the ground, the direction it takes off in and so on.
These humble little sensors are starting to get into everything – boxing gloves, ski boots, barbells, smart-balls to name a few – tracking motion with a pretty decent degree of accuracy. More than anything, they’re really driving innovation in sports tech.
The final category of sensor we’re going to mention is the strain gauge.
These are pretty specific to cycling – at least, from the perspective of this article. By accurately measuring the force that the cyclist is putting through certain critical points on the bike frame, it’s possibly to accurately and consistently measure the power that they’re generating.
Currently, we’re seeing performance sensors like the ones I’ve described above spreading into new sports – or capturing additional data in the sports where they’re already applied.
This leads to another problem: capturing all this data can quickly be overwhelming without the intelligence to analyse it and turn it into actionable insights.
This gap is being filled by a new generation of sports software – the subject of the next post in this series.
If you enjoyed this piece on performance sensors, don’t forget to check out my intro post to this series about my journey up to this point training for Ironman UK.