Title :
Control strategies for visual tracking from a moving platform
Author :
Giesbrecht, J. ; Goldsmith, P. ; Pieper, J.
Author_Institution :
Defence R&D Canada, Ralston, AB, Canada
Abstract :
This paper outlines the control algorithms used in a pan/tilt/zoom (PTZ) tracking system for an Unmanned Ground Vehicle (UGV), implemented as part of a computer-vision based autonomous convoying system. The system relies upon a Linear Quadratic Gaussian controller to keep the target centered in the camera´s field of view using the pan and tilt degrees of freedom. A novel zoom controller, based upon the current target size and various measures of system noise, controls the camera focal length to maintain an appropriate image size for visual tracking, despite changing distances between the leader and follower vehicles.
Keywords :
cameras; linear quadratic Gaussian control; mobile robots; path planning; remotely operated vehicles; road vehicles; robot vision; camera focal length control; computer-vision based autonomous convoying system; leader-follower vehicles; linear quadratic Gaussian control; moving platform; pan/tilt/zoom tracking system; unmanned ground vehicle; visual tracking; zoom controller; Cameras; Control systems; Kalman filters; Noise; Pixel; Target tracking; Vehicles; PTZ control; Unmanned ground vehicles; visual tracking;
Conference_Titel :
Electrical and Computer Engineering (CCECE), 2010 23rd Canadian Conference on
Conference_Location :
Calgary, AB
Print_ISBN :
978-1-4244-5376-4
Electronic_ISBN :
0840-7789
DOI :
10.1109/CCECE.2010.5575246