Section: New Results

Highlights of the Year

Link between performance and risk of injury

Participants : Richard Kulpa [contact] , Benoit Bideau, Michaël Ropars.

In our previous biomechanical analysis of the tennis serve, we have demonstrated that the energy flow is a pathomechanical factor, that means that it can increase joint constraints (and thus risk of injury) while not increasing performance. Nevertheless, the definition and evaluation of energy flow is still a complex scientific challenge. We have proposed to compare the energy flow during the serve between injured and non-injured tennis players by investigating the relationships between the quality and magnitude of energy flow, the ball velocity and the peaks of upper limb joint kinetics [11] . The results showed that ball velocity increased and upper limb joint kinetics decreased with the quality of energy flow from the trunk to the ‘hand+racket’. Injured players showed a lower quality of energy flow through the upper limb kinetic chain, a lower ball velocity and higher rates of energy absorbed by the shoulder, the elbow than non-injured players. These findings imply that an effective energy flow through the kinetic chain by using a proper serve technique is necessary for reducing overuse joint injury risks.  

ACM SIGGRAPH Course on crowd simulation

Participant : Julien Pettré [contact] .

Crowds for entertainment or safety applications purposes are most of the time simulated using microscopic algorithms. In contrast with other types of approaches, microscopic simulators are able to generate continuous and smooth trajectories for individual agents. They are based on models of local interactions between agents. The crowd motion result form the combination of all local motion and interactions. The fact that the resulting crowd motion is emergent makes difficult anticipating the simulation results. Many motion and interaction models have been proposed to design a plethora of simulation algorithms: force-based models, rule-based models, coupled or not with flow-based models, etc. Each type of interaction models will actually result into specific crowd motions as well as agents trajectories. Unfortunately, not all have the desired properties: oscillations, jerky trajectories, residual collisions or deadlocks are often observed in simulations. From this point of view, the course [28] presents the many recent progresses in crowd simulation algorithms since the introduction of velocity-based algorithms , as well as the impact on the level of realism and the visual quality of simulated crowd motions. It also presents the impact on various kind of applications.

Immersive basketball playing

Participants : Franck Multon [contact] , Alexandra Covaci, Anne-Hélène Olivier.

The paper has received the best paper award of the ACM VRST 2014 Conference in November 2014. This paper addressed the problem of perception of distances in immersive environments when dealing with precision distant tasks, such as basketball free throw. The work has been done in collaboration with University of Brasov in Romania, thanks to the FP7 VISIONAIR infrastructure project. The main results of this work tend to show that third person perspectives enabled subjects to perform the task with similar movements than in real world, compared to first person perspective. Third person perspective consists in placing the camera at a different place from the eye point of view, as in many videogames. On the opposite first person perspective consists in place the camera at the place of the user's eyes in scale 1:1, as if the user was colocalized in the virtual environment. We also demonsrated an adaptation to the task in immersive environments, which is a key information for future development of training methods using VR. We have been invited to submit an extended version of the paper to the IEEE Computer and Graphics journal for 2015.

Best Paper Award :

[21] Third Person View And Guidance For More Natural Motor Behaviour In Immersive Basketball Playing in ACM Symposium on Virtual Reality Software and Technology VRST.

A. Covaci, A.-H. Olivier, F. Multon.