Bidirectional invariant representation of rigid body motions and its application to gesture recognition and reproduction

Abstract

In this paper we propose a new bidirectional invariant motion descriptor of a rigid body. The proposed invariant representation is not affected by rotations, translations, time, linear and angular scaling. Invariant properties of the proposed representation enable to recognize gestures in realistic scenarios with unexpected variations (e.g., changes in user's initial pose, execution time or an observation point), while Cartesian trajectories are sensitive to these changes. The proposed invariant representation also allows reconstruction of the original motion trajectory, which is useful for human-robot interaction applications where a robot recognizes human actions and executes robot's proper behaviors using same descriptors. By removing the dependency on absolute pose and scaling factors of the Cartesian trajectories the proposed descriptor achieves flexibility to generate different motion instances from the same invariant representation. In order to illustrate the effectiveness of our proposed descriptor in motion recognition and generation, it is tested on three datasets and experiments on a NAO humanoid robot and a KUKA LWR IV+ manipulator and compared with other existing invariant representations.

Invariant Descriptors of Motion and Force Trajectories for Interpreting Object Manipulation Tasks in Contact

“…Invariant descriptors can be categorized in many types [18] depending on the features that are extracted from the curve. Many have been proposed, both for point trajectories [16], [19], [20], [21], [22] and for rigid-body trajectories [23], [24], [25], [26], [27].…”

Identification of Common Force-based Robot Skills from the Human and Robot Perspective

“…Many frameworks for action or skill classification use a predefined set of actions or skills, e.g. [8], [9]. These skills are usually task specific and have not been proven to be interpretable by humans.…”

Generalizing demonstrated motion trajectories using coordinate-free shape descriptors

“…Motion models constructed from the shape descriptors thus become more generally valid. Shape descriptors have already proven their use in motion recognition [41,26,45], reducing the search space during classification by eliminating unwanted coordinate variations, and hence resulting in an improved recognition rate. In this paper, the goal is to apply the shape descriptor approach to the problem of motion generalization.…”

“…The first contribution is the mathematical formulation of the generative equations that define the rigid-body trajectory as a function of a generic coordinate-free shape descriptor. This compact, generic set of equations is elaborated for two types of existing shape descriptors: the extended Frenet-Serret invariants [41,26] and the screw axis invariants [16].…”