When GPS fails, movement data disappears. A Norwegian developer wants to solve the problem.

19 October 2026

GPS has made movement analysis a central part of modern sport. But underwater, indoors and in challenging terrain, satellite signals can disappear or become unreliable. Norwegian developer Sirar Salih is now developing an algorithm to reconstruct movement paths using existing sensors. The technology is still a prototype, but raises an important question: Can we obtain more complete and reliable movement data where GPS does not work?

GPS has made movement analysis a central part of modern sport. But underwater, indoors and in challenging terrain, satellite signals can disappear or become unreliable. Norwegian developer Sirar Salih is now developing an algorithm to reconstruct movement paths using existing sensors. The technology is still a prototype, but raises an important question: Can we obtain more complete and reliable movement data where GPS does not work? <h2>A swim revealed the problem</h2> It started with a swim. When software developer Sirar Salih came out of the water and checked his training session, Strava showed a straight line between his starting point and the place where he returned to shore. The Garmin watch had been unable to record the actual swimming route underwater. The experience became the starting point for a development project now called Trajectory Positioning System (TPS). The project was covered by <a href="https://www.shifter.no/nyheter/ble-lei-av-rette-streker-i-strava-na-vil-han-sla-gps/478786">Shifter on 13 October 2026</a>. Salih is developing the technology through Oslo-based company <a href="https://wintrytech.com/">Wintry Tech</a>, alongside a full-time job. The aim is to reconstruct movement paths in situations where GPS signals are missing or unreliable. <h2>Reconstructing movement using sensors</h2> GPS uses satellite signals to calculate position. The technology works well in many outdoor situations, but has clear limitations. In practice, satellite signals do not penetrate water, and positioning can become difficult indoors or in environments with obstacles that block or reflect the signals. TPS approaches the problem differently. Instead of relying on continuous satellite signals, the algorithm uses data from accelerometers and gyroscopes. These sensors measure acceleration and rotation, among other things, and are already found in many smartwatches, mobile phones and other electronic devices. By analysing the sensor data, TPS attempts to reconstruct the movement path during periods when GPS positions are unavailable. According to Shifter, the algorithm is approximately one megabyte in size. It is a small software component, but its size alone says little about computing requirements, energy consumption or how well the solution can be integrated into existing products. For now, analysis is performed afterwards. Real-time analysis is a future development goal. <h2>Tested against research data, but not independently validated</h2> Salih has used publicly available research datasets from institutions including NTNU, the University of Oxford, Carnegie Mellon University and ETH Zürich. The datasets cover different environments and applications, including robots, underwater vehicles and movement in challenging conditions. According to Shifter, the developer reports results of up to 97 per cent accuracy compared with reference data. However, this is Salih's own claim. The research institutions have not approved, certified or independently validated TPS through these datasets. Nor is there publicly available documentation explaining how the accuracy metric is defined or how the algorithm performs during actual sporting activities. Assessing the technology will therefore require examining the size of positioning errors, how errors develop over time and how the system handles rapid changes of direction. <h2>A relevant problem for Norwegian sport</h2> Accurate movement data matters in a wide range of sports. In cross-country skiing and biathlon, position and speed data are used to analyse movement along the course. In swimming, GPS-based solutions can have limitations when the sensor is underwater. Indoor sports often use technologies other than GPS, including camera-based systems and local positioning solutions. What these applications share is a need for reliable data about where and how athletes move. When parts of a movement path are not recorded, analyses can become less complete. If these gaps can be reconstructed with sufficient precision, this could create new opportunities for training and performance analysis. There is currently no documented evidence that TPS meets the requirements of such sporting applications. <h2>A familiar technical problem with new possibilities</h2> The principle behind TPS belongs to an established technological field, often referred to as inertial navigation. Using measurements of acceleration and rotation, algorithms can estimate how a person or object moves. The challenge is that small measurement errors accumulate over time. This can lead to growing differences between the calculated and actual position. Researchers and technology companies have therefore spent many years working on methods that combine sensor data, statistical models and machine learning to reduce such errors. The Norwegian School of Sport Sciences has also researched positioning systems for sport. A <a href="https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2018.00115/full">scientific study published in 2018</a>, co-authored by Matthias Gilgien among others, examined the accuracy of a commercial positioning system in indoor team sport. The study showed that installation and measurement conditions could significantly affect the results. This illustrates why documented measurement accuracy is essential when positioning data is used to analyse performance. TPS is therefore not the first technology to address positioning without GPS, but represents a new Norwegian development project in a relevant field of research and technology. <h2>What could this mean for sports technology?</h2> What makes TPS interesting is not necessarily its potential to replace GPS. The opportunity lies instead in maintaining or reconstructing movement information when satellite positioning is unavailable. If the algorithm proves sufficiently accurate, it could potentially enable more continuous movement tracking and reduce data loss. Possible applications include post-session training analysis, tracking movement in challenging environments and improving existing products that combine GPS with other sensors. In the longer term, real-time applications could also be worth investigating. But these applications require further development and validation. In particular, performance must be documented using sensor data from actual sports equipment and under realistic movement conditions. <h2>From prototype to sports technology</h2> Salih's project illustrates how software development can create new opportunities within established sensor technologies. Many sports technology products already contain the sensors on which TPS is based. The question is whether better algorithms can extract more information from these sensors when GPS is unavailable. For Norwegian sports technology, this could be an interesting area for collaboration between technology companies, research communities and sports organisations. The next step for TPS will be to document its performance through independent testing and identify specific applications where the solution adds measurable value compared with existing alternatives. For Norwegian Sport Tech, the interesting prospect is more complete and reliable movement data, including in situations where GPS alone is not enough. It is still too early to know whether TPS can deliver this in practice. But the problem the technology is trying to solve is highly relevant to the future of sports technology. Illustration: An open-water swimmer at sunset. The image does not show TPS or testing of the technology.