Multi-Sensor System based on Unscented Kalman Filter

J.R. Gómez, J. Elías Herrero, C. Medrano, and C. Orrite (Spain)


Tracking Algorithms, Unscented Kalman Filter, Multi Sensor.


Unscented Kalman Filter (UKF) is a very popular tracking algorithm which provides a way of processing non-linear and non-Gaussian models. In this paper we propose a mod ified UKF to extend its applications to multi-sensor scenes, improving thus the global result. The combination of sev eral independent sensors increases the precision and robust ness of our tracking system, making possible to solve dif ficult situations, such as occlusions or serious noise. We approach this modification to multi-camera tracking appli cations, like video surveillance or sport performance im provement.

Important Links:

Go Back