TY - Generic T1 - Tracking Multiple Objects with Kalman Filters, part II Y1 - 2006 A1 - Mikael Lindeborg ED - Marta Marron KW - Erasmus KW - Kalman filters KW - Multi-Object Tracking KW - Real-time implementation KW - TFC AB -

In this report the implementation of a multiple object tracking algorithm is described. The algorithm is part of the obstacle avoidance system in an autonomous robot. The measurement vector used to achieve the tracking task comes from a stereo-vision system that detects objects in the robot’s environment [1]. The algorithm uses the probabilistic Kalman filter (KF) to estimate the position and movement of different objects in the scene. One filter is used for each object to track. An algorithm for associating the data in the measurement vector to different objects is described. A validation process that the tracking algorithm uses to reduce the noise included in the measurement vector is also described.

PB - Department of Electronics CY - Alcala de Henares (SPAIN) U1 - Master in Electronics U2 - Escuela Politecnica Superior U3 - University of Alcala ER -