The vital role of GPS accuracy in wearable tech
As the streets of London filled with the footfalls of Sunday's marathon runners, we've delved into a crucial hot topic: the accuracy of GPS measurements of today's wearables. Whether you're an athlete aiming for a personal best or a first-time marathoner seeking to conquer the distance, the precision of GPS data—capturing an accurate distance, time, and pace—holds significant importance.
Whilst race courses are measured by the shortest possible route (SPR) in accordance with certified guidelines, achieving precise GPS measurements for marathon and daily fitness running poses a significant challenge. Over the span of 26.2 miles, minor deviations caused by GPS accuracy, especially in urban environments - result in a recorded distance that often exceeds the official marathon length.
In this blog, we've unravelled the reason for such measurement inaccuracies and considered the innovative solutions that are emerging within the industry to address them. Before we do, let's recap on the considerable benefits of GPS wearables in supporting distance, pace and performance.
Contributing significantly to a healthier lifestyle, wearable devices:
Distance tracking: Accurate GPS measurements allow runners to track their distance, providing them with valuable data for monitoring their training progress and setting distance-based goals.
Pace monitoring: They enable runners to monitor their pace in real-time, helping them maintain a consistent speed throughout their run or adjust their pace as needed to meet their training objectives.
Performance analysis: By recording accurate GPS data, runners can analyse their performance post-run, examining factors such as pace variations, elevation changes, and split times. This analysis helps them identify areas for improvement and adjust their training regimen accordingly.
However for all this data to have meaningful impact, consumers need the data to be a source of truth. In reality, runners that took part in Sunday’s marathon would have recorded more than the marathon’s 26.2 miles, with one runner who was happy to share their data recording over 26.6 miles on their fitness watch - an increase of over 0.4 miles in terms of the race’s actual distance. Further to inaccuracy in activity metrics, the positioning is often subject to large error, especially in highly dense urban environments - see below as the marathon runner’s path travels through water and high rise buildings!
This Strava route map demonstrates GPS inaccuracies around Canary Wharf
Why GPS accuracy varies on wearable devices
Matteo Ciprian, FocalPoint’s Machine Learning Engineer, helps to explain why there are discrepancies in GPS data, especially in confined spaces, city centre environments or under dense trees. “The design of antennas in wearables leads to poor satellite reception which can lead to compromised positioning accuracy. In order to get ‘GPS ready’, as quickly as possible, a typical wearable starts with only a small portion of the available satellites. This means, your watch may think that you are tens or even hundreds of metres away from your true position, giving you an incorrect read of distance travelled and incorrect line on a map. That first GPS fix is critical to the calculation of your route, so a bad first fix from a device will ruin the first few hundred metres of your run. This is because your device struggles to catch up and smooth out the differences between the first fix and the subsequent ones. GPS is further challenged in wearable devices by the high dynamics that they experience. As runners swing their arms and bounce along the road, this introduces significant doppler (frequency shift) to the satellite signals, making tracking satellites even more difficult, further delaying ‘GPS ready’ and introducing errors in positioning accuracy.”
In summary, the challenges faced by wearable devices manifest as:
‘GPS Ready’ AKA Time to first fix (TTFF): First position is often poor due to low number of satellites and can introduce early error to activity metrics
Poor accuracy during activity: due to platform hardware constraints and high dynamics, GPS positioning struggles to maintain accuracy during running.
Effective strategies OEMs need to consider to overcome these GPS inaccuracies
Addressing these discrepancies involve wearable OEMs to consider their most effective strategy for boosting accuracy:
Multi-frequency GNSS: wearable OEMs are increasingly offering dual-band GNSS in their high-end running watches. These use an additional set of GNSS signals known as L5, alongside the previous generation, L1. Dual band receivers have the potential to deliver higher accuracy, but have further size, weight and power requirements, and come at an additional cost.
Antenna design: no-one wants a huge weight on their wrist when exercising. This has a direct impact on the antenna design and consequently means the reception for GNSS signals is highly compromised versus, for example, a vehicle with a large external antenna. OEMs choosing to place further priority on the antenna design, must consider the trade off against other space-saving consumable devices.
Software & Sensor Fusion: Software has the advantage of being a highly scalable and cost effective solution. There are a few options open to OEMs. Combining sensor fusion with GNSS boosts areas of poor satellite measurements with additional data from sensors like accelerometers, gyroscopes and magnetometers. Combining multiple data sources in this way can produce a significant boost to positioning accuracy. However this approach is not without limitations. Areas with poor GPS, such as highly urban environments, can become unusable - even when fused with sensor outputs.
In contrast, FocalPoint’s approach used machine learning and sensor fusion together with Supercorrelation technology, to enable wearable devices to track only the line of sight signals and boost receiver sensitivity. This effectively mitigates the poor sensitivity that comes from highly constrained antenna design.
No matter your level of running - highly accurate GPS measurements are essential for race preparation, enabling effective race strategies, including setting realistic performance goals. In our next blog, Matteo Ciprian explains what's happening with the GPS and inertial navigation data.
If you’d like to know more about our Supercorrelation™ technology and its application to the wearables sector, please get in touch.