For this situation, the GPS signal very weak propagate in an indoor envirroment. Normally, the researchers use the signal from another radio signal soures, for example WiFi signal, X-bee. Those souse can use to be sub-system for positioning, however, how to integate the systems between indoor and outdoor still a main prolem for local based service.  Good luck

Generally, inertial sensors can be more useful.  There are very inexpensive systems available for humans, because the dynamic range of humans is very limited, compared to machines. 

Inertial methods can also be used to integrate other data (RF systems) and to help merge ensemble methods.

In nearly all cases "how it works" involves both signal processing of the sensor data, and filtering (e.g., Kalman filters) to estimated position, velocity, acceleration, or angular data.

GPS is a good example - the signal coming down from space is a clock.  Your local system must navigate by comparing the time signals from several satellites.  There is no such thing as direct navigation from a GPS signal. 

The only systems which "directly" observe position, etc, are imaging systems which often require the athletes to wear reference points or reflectors.  Computer processing can do what we used to call cinetheodolite calculations.  In theory this can be done without signal processing or filtering.  In practice...??? 

Indoor positioning and indoor gnss are two different things.

One example of indoor positioning is http://www.locata.com/2011/09/locata-develops-indoor-positioning-technology-for-industrial-and-warehouse-applications/

On the other hand indoor gnss relies on the reflected signals available inside a building. These signals can not be used to compute the precise position of the receiver. It can only be used for a coarse position or to keep your GPS alive.