Imitation of human body poses and hand gestures using a particle based fluidics method

Download

index.pdf

Date

2012

Author

Tilki, Umut

Metadata

Show full item record

Item Usage Stats

261
views

142
downloads

In this thesis, a new approach is developed, avoiding the correspondence problem caused by the difference in embodiment between imitator and demonstrator in imitation learning. In our work, the imitator is a fluidic system of dynamics totally different than the imitatee, which is a human performing hand gestures and human body postures. The fluidic system is composed of fluid particles, which are used for the discretization of the problem domain. In this work, we demonstrate the fluidics formation control so as to imitate by observation initially given human body poses and hand gestures. Our fluidic formation control is based on setting suitable parameters of Smoothed Particle Hydrodynamics (SPH), which is a particle based Lagrangian method, according to imitation learning. In the controller part, we developed three approaches: In the first one, we used Artificial Neural Networks (ANN) for training of the input-output pairs on the fluidic imitation system. We extracted shape based feature vectors for human hand gestures as inputs of the system and for output we took the fluid dynamics parameters. In the second approach, we employed the Principal Component Analysis (PCA) method for human hand gesture and human body pose classification and imitation. Lastly, we developed a region based controller which assigns the fluid parameters according to the human body poses and hand gestures. In this controller, our algorithm determines the best fitting ellipses on human body regions and human hand finger positions and maps ellipse parameters to the fluid parameters. The fluid parameters adjusted by the fluidics imitation controller are body force (f), density, stiffness coefficient and velocity of particles (V) so as to lead formations of fluidic swarms to human body poses and hand gestures.

Subject Keywords

Posture., Imitation., Robots, Robots, Robots, Computational fluid dynamics., Neural networks (Computer science).

URI

http://etd.lib.metu.edu.tr/upload/12615140/index.pdf
https://hdl.handle.net/11511/22056

Collections

Graduate School of Natural and Applied Sciences, Thesis

Suggestions

OpenMETU
Core

Imitation of basic hand preshapes by fluid based method: fluidics formation control Tilki, Umut; Erkmen, İsmet; Erkmen, Aydan Müşerref (2011-01-01) In this paper, a new approach is developed, avoiding the correspondence problem caused by the difference in embodiment between imitator and demonstrator in imitation learning. In our work, the imitator is a fluidic system of dynamics totally different than the imitatee, which is a human performing hand gestures. The fluidic system is composed of fluid particles, which are used for the discretization of the problem domain. In this work, we demonstrate the fluidics formation control so as to imitate by observ...
Imitation of Basic Hand Preshapes by Fluid Based Method: Fluidic Formation Control Erkmen, Aydan Müşerref; Erkmen, İsmet (2009-11-05) In this paper, a new approach is developed handling the correspondence problem due to the difference in embodiment between imitator and demonstrator in imitation learning. In our work the imitator is a fluidic system of dynamics totally different than the demonstrator which is human hand gestures. In this work we demonstrate the fluidic formation control so as to generate basic hand preshaping features. Our fluidic formation control is based on Smoothed Particle Hydrodynamics (SPH), which is a particle base...
Implementation of a closed-loop action generation system on a humanoid robot through learning by demonstration Tunaoğlu, Doruk; Şahin, Erol; Department of Computer Engineering (2010) In this thesis the action learning and generation problem on a humanoid robot is studied. Our aim is to realize action learning, generation and recognition in one system and our inspiration source is the mirror neuron hypothesis which suggests that action learning, generation and recognition share the same neural circuitry. Dynamic Movement Primitives, an efficient action learning and generation approach, are modified in order to fulfill this aim. The system we developed (1) can learn from multiple demonstr...
Robotic system design for reshaping estimated human intention in human-robot interactions Durdu, Akif; Erkmen, İsmet; Erkmen, Aydan Müşerref; Department of Electrical and Electronics Engineering (2012) This thesis outlines the methodology and experiments associated with the reshaping of human intention via based on the robot movements in Human-Robot Interactions (HRI). Although works on estimating human intentions are quite well known research areas in the literature, reshaping intentions through interactions is a new significant branching in the field of human-robot interaction. In this thesis, we analyze how previously estimated human intentions change based on his/her actions by cooperating with mobile...
Momentum transfer continuum between preshape and grasping based on fluidics Özyer, Barış; Erkmen, İsmet; Erkmen, Aydan Müşerref; Department of Electrical and Electronics Engineering (2012) This dissertation propose a new fluidics based framework to determine a continuum between preshaping and grasping so as to appropriately preshape a multi-fingered robot hand for creating an optimal initialization of grasp. The continuum of a hand preshape closing upon an object that creates an initial object motion tendency of the object based on the impact moment patterns generated from the fingers is presented. These motion tendencies should then be suitable for the proper initiation of the grasping task....

Citation Formats

U. Tilki, “Imitation of human body poses and hand gestures using a particle based fluidics method,” Ph.D. - Doctoral Program, Middle East Technical University, 2012.