When tracking workload in professional sports, it is crucial for any staff to have a reference model, i.e. the match data. To reach this goal, the use of GPS devices in the official games is one of the alternatives. Still, its effectiveness –related to GPS accuracy– largely depends on the environment. Modern stadiums are equipped with high stands, built very close to the playground, or even with partial or total coverage: all characteristics that hinder the visibility of the satellites and make the GPS data less accurate.

Exelio’s CEO, Valerio Roberti, helps us contextualise and narrow the problem by answering some questions.

1. GPS lack of accuracy in modern stadiums: how can you deal with this issue?

There are procedures to improve the accuracy of the GPS data:

• enable more constellations to increase the number of satellites visible in the sky; 
• take advantage of the IMU data when the GPS reception loses quality. However, this is a solution that can only work for short periods of time, because the data of the inertial unit tends to derive over time. This compensation usually occurs when the GPS data lose quality but is only effective for a few seconds.

No GPS, regardless of brand, can work in modern stadiums at the same accuracy they have on outdoor training pitches.

2. How else can you track match data?

Currently, video tracking is widely used. The technology has improved over time, and the data has become accurate enough to be finally compared with GPS. Still, some precautions must be taken to facilitate this comparison. Even accurate raw data from video-tracking does not guarantee comparability with GPS final data. An intervention through a filter is necessary, according to the filter used by the GPS system, in order to obtain comparable final data.

3. What are the advantages and limitations of video-tracking?

The main advantages are that with just a system you can monitor both teams, as well as track the ball, and the athletes don’t have to wear anything.

On the other hand, for those who make large use of live data, video tracking does not have the same accuracy as the final data, which is only provided after processing (human or AI intervention). It can last several hours. This processing, which is based on data cleaning, finally makes the information possible to compare with that collected via GPS during training.

4. Is there any convincing alternative?

Some tools, which are based exclusively on the accelerometer, empirically measures some parameters such as momentum, or specific type of events such us impacts, jumps. This approach does not fully convince us, because it does not allow to have accurate data nor to have metrics with a solid scientific basis.

A valid alternative is the LPS system. It is usually based on UWB (Ultra-Wideband) technology, which requires some antennas to be installed on the sidelines. Athletes wear a device that sends a signal to these antennas. Starting from the time of flight of this signal, a central unit can calculate an high-accurate position of the athlete time by time and derive the different metrics. Our LPS system can achieve a position accuracy of 10 cm.

5. At what sampling frequency does the LPS work?

It works at 25 Hz: a frequency that we consider more than suitable (higher than needed) to monitor human locomotion, even in sports with frequent speed variations, such as football. Regardless of the technology (GPS or LPS), we experienced that devices working at higher frequencies do not add any useful information, and vice versa this will only result in a repetition of the same data for more samples, with a bad impact on the data storage and the time of data transfer. 

6. Does the LPS really work in modern stadiums?

In addition to some installations already carried out in indoor arenas, for other sports and facilities of different sizes, we have already tested LPS in football stadiums. The results confirm what we expected. We are working to carry out an official validation. 

The positioning of the antennas depends on the structure of the environment, the presence of any obstacles and the visibility of spectators and sponsors. This is one of the challenges: finding a positioning of the antennas that do not cause trouble but is effective for the perfect functioning of the system. Field size is not a constraint. In the detail of the tests we carried out, the correct coverage of the system was verified by running in all areas of the field.

Test of LPS (dark green) vs GPS system inside a stadium.

7. What are the LPS technology limits if compared to video tracking?

Compared to video tracking, the limitations are related to the fact that the player has to wear a device. However, the device can reach very small sizes, as in our new LPS tag model that will be released on the market in the spring, which will have dimensions of 52 x 33 x 14 mm. Secondly, the opponent team’s players would not be tracked, if they do not equip themselves with the tags.

Furthermore, ball tracking is missing for those who need a tactical analysis. There are companies that manage the tracking of the ball by inserting a chip inside it. This is a technically feasible solution, but we see limitations in terms of system operability and scalability (e.g.: agreements with the various manufacturers, which are different for each league). We are evaluating a hybrid system with video tracking of the ball.

Finally, in the market we can find many solutions and a dynamic scenario due to quick technological developments. For this reason, we advise each potential user to make an assessment according to their specific needs and consequently choose the most suitable solution.

Interview with Valerio Roberti, Exelio CEO.