Probabilistic Object Tracking using a Range Camera

TL;DR

This paper presents a probabilistic method for real-time 6-DoF object tracking using a range camera, explicitly modeling occlusions and employing a Bayesian network with particle filtering for robust pose estimation during manipulation.

Contribution

It introduces a novel Bayesian network approach with occlusion modeling and Rao-Blackwellised particle filtering for improved object tracking during manipulation.

Findings

Accurately tracks object pose in real-time during manipulation.

Robustly handles self-occlusions and environmental occlusions.

Demonstrates effectiveness with human and robotic manipulation.

Abstract

We address the problem of tracking the 6-DoF pose of an object while it is being manipulated by a human or a robot. We use a dynamic Bayesian network to perform inference and compute a posterior distribution over the current object pose. Depending on whether a robot or a human manipulates the object, we employ a process model with or without knowledge of control inputs. Observations are obtained from a range camera. As opposed to previous object tracking methods, we explicitly model self-occlusions and occlusions from the environment, e.g, the human or robotic hand. This leads to a strongly non-linear observation model and additional dependencies in the Bayesian network. We employ a Rao-Blackwellised particle filter to compute an estimate of the object pose at every time step. In a set of experiments, we demonstrate the ability of our method to accurately and robustly track the object pose in real-time while it is being manipulated by a human or a robot.