Taking bio-mechanics out of the laboratory

We represent the human being as a mechanical system. Instead of having steel parts like in a machine, we put bones. Instead of motors, we use muscles. Instead of links with bearings, we use ligaments, bone surfaces, and so on. We then apply the equations of mechanics to this system to assess how the human functions in interaction with its environment, or with an object. This could be a piece of sports equipment, a tool for work, or a medical device to assist in a disabled situation.

This multidisciplinary team^[1] is the successor to MimeTIC, which for 13 years explored innovative methods for modeling human movement.

An illustration of the complexity of the phenomena at work: pole vaulting. “The pole vaulter runs. He stores kinetic energy as he runs. Then he plants his pole. He bends it. He then transfers his kinetic energy to the pole in the form of elastic potential energy. He rises into the air. He puts his muscles to work. And the pole gives him back the energy it had stored. Many phenomena need to be taken into account: how force is transferred, how energy is transferred, how to maximize these transfers to achieve the highest possible jump, etc.” explains Charles Pontonnier, head of ComBO.

Body surface measurements

This type of investigation has to contend with a central problem: “we can't install sensors inside the body to measure the internal stresses we're trying to assess. We only have external sensors. When we measure a person's muscular activity, it's only on the surface. We then need methods to estimate what's going on inside on the basis of what we're able to measure on the surface. But their validation is proving complex.”

Scientists^[2] have therefore invested a great deal of time in the laboratory to build these methods and also to render the interaction in extended reality. This has led to a host of innovations, particularly in sport. In soccer, for example, goalkeepers can now train to better manage their defense thanks to simulation. In cycling, as part of the European Sharespace project, a new breakaway training system will even attempt to saddle a peloton  whose riders will not all be physically present on the circuit.

Two technology platforms

For all this work, the ComBO team relies on two rather unique pieces of equipment. Firstly, Immersia, a virtual reality room that can be used to set up very large-scale experiments. Secondly, Immermove, a platform that includes, among other things, an entire gym dedicated to motion capture.

We've reached a turning point where our methods are now mature. We want to get them out of the laboratory and into the hands of people in the field, whether they be sports coaches, ergonomists or carers.

We're no longer working under laboratory conditions. A lot of things change in relation to the person's real environment. This generates very different constraints. We become dependent on external conditions. We can't equip people in the same way. The data is less good, less numerous, noisier.”

For the time being, however, while laboratory models work well, they still require the implementation of complex methods, expertise on the part of the person deploying them and, above all, lots of measurements. Question: “If we move on to data with a lower level of information and precision, how do we solve our equations? How can we ensure that they still make sense ?”

The solution could come from Machine Learning. “There's a good chance that we can exploit the information we have on statistical models of humans to supplement less precise measurements with precise information. That's our aim: to augment the data we obtain in the field with data from the laboratory or larger databases to build representative models of the person being studied from highly degraded measurements. This type of approach requires modern machine learning methods, which we exploit in the team.”

Highly applicative positioning

The team prides itself on its highly applicative approach. “Whether it's sport, ergonomics in the workplace or disability, it's in the field that we draw the issues that fuel our scientific questioning. And it's to the field that we then return with adapted answers.”

In the clinical field, for example, ComBO is working with the Rainbow robotics team and the Saint-Hélier rehabilitation center in Rennes on an upper-limb exoskeleton for patients with multiple sclerosis or who have suffered a stroke. “These people can sometimes only mobilize their arms with very low levels of force. The exoskeleton can improve the situation. But it has to respond in a way that's best suited to each individual's condition. So there's an interest in modeling humans in this type of situation to build up an image of their effort-generating capabilities.” This research is being carried out as part of the Inria Exploratory Action^[3]MusMapS.

The scientists were involved in the development of another exoskeleton, this time for butchers in the food processing industry. This profession is highly exposed to musculoskeletal disorders (MSD). Led by LAB4i, the Ovalt group's innovation subsidiary, with the support of the Cooperl group, the Exoscarne 2.0 collaborative project therefore aimed to find solutions to reduce the level of manual effort. “In this context, we were asked to characterize the task, understand the bio-mechanics of the butcher when cutting a piece of meat, and then evaluate the interaction with the exoskeleton. Evaluate and analyze the beneficial and deleterious effects of the system developed. We intervened upstream to define the specifications in the biomechanical sense, and downstream to evaluate the solution.” The device is now being tested on a cutting line. It's worth noting that some of the team's researchers have also helped launch Moovency, a start-up which is also involved in MSD prevention.

Helping prepare athletes for the 2024 Olympics

On the sports front, it was the Paris Olympic Games that recently mobilized researchers through three Priority Research Programs (PPR): BEST Tennis on players' return of serve, Revea on optimizing athlete performance through the use of virtual reality, and Neptune, on pool swimming. “All this has given rise to a transfer of scientific results and their direct use for top-level sport through sports federations.”

_{^[1]Combo is a joint Inria, Université de Rennes, Université de Rennes 2 and ENS Rennes project-team within the Mouvement, Sport, Santé (M2S) laboratory at Université Rennes 2 and the Irisa laboratory.}
 

_{^[2]The permanent members of ComBO are : Franck Multon, Nolwenn Fougeron, Benoît Bideau, Nicolas Bideau, Armel Cretual, Diane Haering, Hugo Kerhervé, Richard Kulpa, Caroline Martin, Guillaume Nicolas, Anthony Sorel, Mathieu Ménard, Nicolas Vignais, Georges Dumont, Ronan Gaugne, Fabrice Lamarche, Charles Pontonnier, Laurent Guillo and Pierre Hellier.}

_{^[3]An Inria Exploratory Action is an internal mechanism to facilitate the emergence of new research themes by giving scientists the means to test original ideas.}