Moving object detection for interception by a humanoid robot
Abstract
Interception of a moving object with an autonomous robot is an important problem in robotics. It has various application areas, such as in an industrial setting where products on a conveyor would be picked up by a robotic arm, in the military to halt intruders, in robotic soccer (where the robots try to get to the moving ball and try to block an opponent's attempt to pass the ball), and in other challenging situations. Interception, in and of itself, is a complex task that demands a system with target recognition capability, proper navigation and actuation toward the moving target. There are numerous techniques for intercepting stationary targets and targets that move along a certain trajectory (linear, circular, and parabolic). However, much less research has been done for objects moving with an unknown and unpredictable trajectory, changing scale as well and having a different view point, where, additionally, the reference frame of the robot vision system is also dynamic. This study aims to find methods for object detection and tracking using vision system applicable for autonomous interception of a moving humanoid robot target by another humanoid robot. With the use of the implemented vision system, a robot is able to detect, track and intercept in a dynamic environment the moving target, taking into account the unique specifications of a humanoid robot, such as the kinematics of walking. The vision system combined object detection based on Haar/LBP feature classifiers trained on "Boosted Cascades'' and target contour tracking using optical flow techniques. The constant updates during navigation helped to intercept the object moving with unpredicted trajectory.
Degree
M.S.
Advisors
Matson, Purdue University.
Subject Area
Robotics|Artificial intelligence|Computer science
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.