As many articles in this blog will discuss GPS technology, let’s start with the basics…
Most practitioners, especially those working in team sports, are exposed to GPS technology. It can seem like a little black box that spits out streams of data; data which we are expected to translate for coaches and players. As a Scientist it is essential we understand the inner workings of the technology we employ, the methodologies behind the numbers and the limitations and errors associated with them.
GPS units can contain a number of components including the satellite receiver, accelerometer, magnetometer, gyroscope and a battery, and here I will focus on the two currently most associated with calculating the data we use (satellite and accelerometer).
Global Positioning Satellite receivers can provide data based on point to point location analysis. There are 24 – 30 satellites orbiting the Earth, originally part of the USA’s military programme and now accessed by many everyday objects, including car sat navs and GPS watches and units for sport and exercise. A receiver needs to lock onto three or more satellites, calculate the distance to each and therefore determine its own location via trilateration.
Satellite visibility affects the signal strength, the more satellites the receiver can ‘see’ the better the quality of data, hence why data can be poor when using GPS in stadiums as the stands block access to multiple satellites. Data quality can also be affected by tall buildings, terrain and foliage, atmospheric conditions, electronic interference and satellite geometry, all of which can cause ‘positional dilution of precision’.
The frequency of the GPS unit dictates how many sampling points per second are registered i.e 5 Hz represents 5 per second. These discreet data points are joined up to estimate the path of the individual so in theory a higher sampling rate should provide data closer to the actual path travelled. However, not all is what it seems when it comes to GPS frequency. Some of the higher advertised frequencies i.e. 15 Hz are not necessarily derived from 15 GPS data points per second but from interpolated data. This is a 10 Hz GPS sampling rate supplemented with accelerometer data in an attempt to improve the reliability of the tracking data.
There is debate regarding the optimal sampling frequency of GPS. With a higher resolution of data collection some points may not represent true human movement plus the error associated with each point could be amplified by a greater sampling frequency. On the other hand more data points may help to improve the collection of smaller, more intense movement such as accelerations and decelerations.
I will be discussing the validity and reliability findings of GPS studies and different sampling frequencies in future posts.
GPS units contain tri-axis accelerometers meaning they measure a composite vector magnitude by calculating the total of proper acceleration across three axes; x, y and z. When a unit is stationary they will measure 0g across the x and z plane but a gravitational force of -1g (or 9.8m/s/s) in the y (vertical) axis due to the Earth’s standard acceleration due to gravity.
I like to think of using accelerometers on rollercoasters as the movement of the body due to g forces is exaggerated:
- When you are thrown from left to right by side on crashes during the bumper cars, the g force acts along the x axis
- The g force from a vertical freefall on Disney’s Tower of Terror or The Detonator at Thorpe Park act in the y axis
- Being driven straight forward into a tunnel on Aerosmith’s limo in Disney’s Rock n Roller Coaster, throws your body backward by the g force acting along the z axis
- The best rollercoasters of course exert g force along all of the axes!!
Accelerometers can be found in a variety of everyday objects, including Wii remotes, pedometers and smart phones. Jonny introduced me to a great app called Accelerometer-Visual that displays tri-axial readings and is a useful tool in understanding the behaviour of an accelerometer. Can you work out how the iPad is being held in the example shown?
Unlike GPS receivers it is widely accepted that the greater frequency of an accelerometer is beneficial given that they are attempting to capture all movement and force going through the unit. There are currently 100 Hz tri-axial accelerometers integrated into the newest GPS units, measuring up to 16G on each axis.
The use of accelerometers in quantifying the forces elicited on the body, through impact and/or collisions, is a very interesting area of research and one I am sure will also be discussed here.
Keep a look out for my future posts delving into the validity and reliability of GPS or if you are ready for more in the meantime look up Jonny’s post ‘Monitoring with GPS: Time to Slow Down?’