Session 2: Artificial Movement, Animation & Imaging

Computational Design of Mechanical Characters

Bernd Bickel

Institute of Science and Technology (IST), Austria

3D printing is considered a disruptive technology with potentially tremendous socioeconomic impact. In recent years, additive manufacturing technologies have made significant progress, in terms of both sophistication and price; they have advanced to a point where devices now feature high-resolution, full-color, and multi-material printing. Nonetheless, they are of limited use, given the lack of efficient algorithms and intuitive tools that can be used to design and model 3D printable content.

My vision is to unleash the full potential of 3D printing technology with the help of computational methods. In my research, I am working to invent and develop new computational techniques for intuitively designing virtual 3D models and bringing them to the real world. In this talk, I will present our efforts towards a general interactive design system for automatically creating the mechanisms that reproduce the desired motions. This will allow novice users, for example, to create complex 3D mechanical characters that are ready for fabrication. Furthermore, I will address the challenge of of designing mechanical automata that can walk in stable and pleasing manners.

Combining low dose irradiation stereoradiography and movement analysis: What could be gained in kinematics?

Hélène Pillet¹, Christophe Sauret¹, Wafa Skalli¹, Morgan Sangeux^2,3,4

1. Arts et Metiers ParisTech, Institut de Biomécanique Humaine Georges Charpak, 151 boulevard de l’Hôpital – 75013 Paris, France

2. Hugh Williamson Gait Analysis Laboratory, The Royal Children's Hospital, Melbourne, Australia

3. The Murdoch Childrens Research Institute, Melbourne, Australia

4. The University of Melbourne, Australia

Functional methods are attractive approaches for the determination of joint center or joint functional axis. However, these methods suffer from a lack of validation with respect to the underlying bones’ anatomy. The EOS medical imaging system offers the opportunity to obtain a 3D subject-specific reconstruction of the skeleton with a low-dose of irradiation. Data fusion between can be used to benefit from this personalized reconstruction in the motion capture environment. This presentation will focus on the quantification of the data fusion accuracy for the hip and the knee joints. As an example of application, the procedure will be used to validate functional knee joint calibration and show the impact of this method on the kinematics of the knee.

Muscle based control for avatar animation : a synergy based approach

Charles Pontonnier^1,2,3, Ana Lucia Cruz Ruiz^1,2, Antoine Muller^1,2, Georges Dumont^1,2

1. IRISA/INRIA MimeTIC, Rennes, France

2. ENS Rennes, Bruz, France

3. Ecoles de Saint-Cyr Coëtquidan, Guer, France

Muscle-based control is transforming the field of physics-based animation through the integration of knowledge from neuroscience, biomechanics, and robotics, which motivate the creation of smarter characters, and most importantly, enhance motion realism. Since any physics-based animation system can be extended to a muscle actuated system, the possibilities of growth are tremendous. However, modeling muscles and their control remains a difficult challenge. This talk aims at presenting the main control approaches used in the animation field for muscle-actuated avatars, with a particular focus on a method currently developed in our research team. This control method is based on muscle synergy extraction and adaptation to drive a direct dynamics simulation. The experimental protocol for synergy extraction and model are first presented, followed by a control method consisting of a series of optimizations to adapt muscle parameters and synergies to match experimental data. An application to a direct dynamics control of a human arm throwing motion is presented. Results show that the motion can be accurately reproduced thanks to the muscle synergy extraction and adaptation to the model, even if challenges remain numerous.

Performances of the avian body design – X-ray video analysis of the bird bipedal locomotion

Anick Abourachid

Muséum National d’Histoire Naturelle, UMR 7179 CNRS, Paris, France

Birds are very diverse, both in the number of species and their ability to live in various environments. Despite this ecological plasticity, the bauplan of these animals is very conservative and corresponds to specializations for flying that marks the bird’s morphology. This specialization necessitates wings and a rigid trunk. Besides these specializations for flying, birds are fundamentally bipeds. This feature, basal for the clade, allows them to move on most substrates without profound modifications of the avian bauplan. The only other strictly biped species, humans, is not capable of such locomotor plasticity. The question we address here is the link between the geometrical features of the bird bauplan and the performances of these bipeds. The comparison of the kinematics of the body parts during different locomotor behaviors: walking, swimming, hopping, and taking off, is a way to assess this question. However, feathers hide the body and X-ray analysis is needed to quantify the movements of the underlying skeletal elements. The trajectories of the bones indicate that two functional modules participate to the locomotor plasticity, the trunk and thigh jointly control the path of the center of mass, and the distal part of the legs propels the system. A kinematic model of the avian limb is compared to the human one. A comparison of the evolutionary trajectories of both bipeds is used to understand the origin of the differences between them. 

From fragments to full-body biomechanics using anatomy transfer

François Faure

Anatoscope, Université Joseph Fourier, INRIA, Grenoble

The reconstruction of a complete anatomy based on fragments or partial data is a hard and tedious task. Anatomy Transfer is a new approach to build complete 3D digital anatomies, by warping a complete reference anatomy to match partial data while respecting anatomical principles. Based on various types of data (surface scans, medical images, dimensions), it automatically produces anatomical models that complement it with plausible extrapolations. The models include mechanical parameters and data and can be simulated using biomechanical simulation software. We present the principles of this approach and comment on its potential and limitations.