The knowledge of the exact position, speed, acceleration and additional parameters of an athlete in real time, can provide useful information to coaches, referees and sport enthusiasts.

Coaches, for example, could track the covered distances, instantaneous and mean velocities or the exact time on field of their athletes. Furthermore, such data would offer easier determination of overall performance and conclusions on energy expenditure.

Besides generating useful data for training purposes, such knowledge would also support tactical decisions depending on different in game situations and physical condition of the athletes, e.g. in football or ice hockey.

Knowing the exact absolute position of an athlete or the relative position to his opponents would support referees in ambiguous situations to make a more reliable decision.

There are several methods to track a single sportsman or a whole team, e.g. GPS and video or radio wave based systems.

The objective of this project is to improve the position estimation of a local positioning system (LPS). This LPS is based on the FMCW radar principle and consists of several fixed base stations and some mobile transponders worn by the athletes.

The intended approach to achieve this aim is to use additional data acquired independently from the LPS. This data is processed by a specific model that describes the movement of an athlete depending on different sports. The secondary position estimation of the model is combined with the initial estimation of the LPS by data fusion techniques. Thus, a more detailed and reliable position estimation is achieved.