Design of an adaptive estimator for tracking football in flight

Estimation of variables in presence of nonlinearity in the trajectory equations as well as observations is a challenge that demands the attention of the researchers in every domain. In recent years there has been great progress in the generic filter design in which the filter undergoes a huge amount of stochastic data processing in real time, a cost that is incurred for getting refined estimates of the system variables. Different variants of Kalman filter and particle filter can be designed. The success of the filter depends heavily upon the ability to formulate the system and measurement equations together with the noise models. In this work, extended and unscented Kalman filter design is conducted for tracking a football during a free kick under some restrictions and assumptions. The design is made adaptive through the incorporation of knowledge regarding the players involved as well as ambient noise statistics. Performance of the filter is tested in a simulation environment through Monte Carlo runs for the proof of concept and it is found to be satisfactory and also leads to important answers regarding realtime integration.