Fewer stumbling blocks

Research at the Locomotion Laboratory

2020/06/19 by

Human movement when stumbling is manifold – and remarkably misunderstood. TU motion scientist have come to a well-founded conclusion that could even help robots to maintain balance.

Professor André Seyfarth, Dr. Maziar Sharbafi and Christian Schumacher (from left to right) adjusting the prototype of a soft-suit exoskeleton.

The Angular Momentum Perturbator (AMP) looks like a backpack. It weighs 16 kilograms, and contains a fast-rotating tied gyro that is suspended from a motorised frame in such a way that torques are transmitted by an additional rotation to the person carrying the „backpack“. The AMP has a simple function: it causes a perturbation of the subject’s posture. The person then has to compensate for this disturbance with their body so as not to fall. For researcher Christian Schumacher and his colleagues at the Locomotion Laboratory at TU Darmstadt, this moment is an opportunity to take a closer look at the specific function of the leg muscles.

Schumacher conducted studies during a research stay at the Biorobotics Lab of the Delft University of Technology (Netherlands), where the AMP was developed. Both research groups were the first in the world to use a system of this kind to investigate the function of the so-called biarticular (two-joint) leg muscles. For the first time, the teams were able to demonstrate through experiments that these muscles responded most strongly when it came to realign the upper body, for instance after a push. An understanding of these processes not only helps to better understand the entire human motor control and to compensate for possible impairments with technology, but also encourages the development of robots that are able to walk more safely and more efficiently on two legs – which is still a tremendous challenge for science.

There are good reasons why the mechanisms of human motor control are not yet fully understood. „The biggest problem in our research is that the body is so diverse,“ explains Schumacher. When we are pushed, stumble or walk on an uneven terrain, our body coordinates a variety of muscles to compensate for the uncertain movements and restore balance. It has a tremendous number of levels of freedom at its disposal – and thus numerous possibilities for moving muscles and joints.

One research hypothesis is that we create a certain overarching representation in the brain. The processed signals transform these decisions into a kind of exact road map for the individual muscles.

„Imagine you want to pick up a pen that‘s on the table in front of you,“ says Schumacher. „You can do it in an infinite number of ways.“ The mechanisms are understood. First, neural impulses are created in the brain, processed in the spinal cord and passed on to the muscles. These impulses create the contractions of the muscles, and with that the motion. At the same time, sensory signals arrive in the spinal cord e.g. from the skin receptors in the fingers as soon as we touch the pen and feel a counterforce. These signals support the control or, in turn, initiate new movements.

Although researchers can observe the movements closely and measure the neural signals from the muscles, they cannot (yet) be predicted. „This is because one has specific aims with every movement that influence the type of movement, for instance to do something efficiently or quickly,“ explains Schumacher. So the goals of a motion are as diverse as the movements themselves. So if researchers are unable to predict how a motion is actually performed based on body structures and hypotheses, it is difficult to explain the roles of individual muscles in the overall complex.

The same also applies to the leg muscles. One of the differences between leg and arm movements is that leg movements are often instinctive. If you reach out for something you plan the movement, but if you are running and then trip, you will have to rely heavily on your reflexes. The response is automatic. The spinal cord takes on a lot of the control by effectively responding directly to the incoming signals and controlling a whole range of muscles at the same time. „One research hypothesis is that we create a certain overarching representation in the brain, for instance: I want to go from A to B with these objectives or under these conditions,“ explains Schumacher. „The signals that are processed in the spinal cord transform these decisions into a kind of precise road map for the individual muscles, which then know exactly what to do and when and how.“

Variability as an advantage

Although this, unlike the picking up of the pen, is an automatic sequence of events, there is the same motion diversity. „There is the same degree of variability between individuals as there is among the actual functions,“ says Schumacher. One may perhaps bend a knee a little more while another makes more use of individual muscles. This variability is an advantage because if an injury prevents a particular function from working, we might be able to use another one – often with almost identical results. This makes us more viable, you could say.

In order to investigate the role of the two-joint muscle, Schumacher and his colleagues divided movements into three basic functions: stance leg function, swing leg function, and upper-body balance. This division into clearly distinguishable functions has enabled researchers to identify the different functional contributions of the two-joint muscles. Among other things, this enabled the scientists to prove, as had been hypothesised, that these muscles were very active in stance leg control. They help us to save energy when pushing ourselves off the ground because they can transfer energy from one joint to the other. This means that we move more efficiently, but also that we have more strength, for instance when hopping.

Now it is a matter energy efficiency is relevant in the field of prosthetics. Most prostheses are still passive today, which is why prosthetic wearers consume up to 60 percent more energy when walking than people without prostheses. Actuated systems that actively support people when walking require a certain battery capacity – but wearers find the resulting greater weight unpleasant. The mechanisms of two-joint muscles can help to do more with less energy.

And finally exoskeletons benefit from these results. They have evolved tremendously in recent years. The cumbersome mechanical scaffolding has now been replaced by soft-suit exoskeletons – textile-based systems that transmit forces using cable pulls. They support people with impaired movement for instance during push-off or leg swing. The TU group has already shown that people with exoskeletons that include the mechanisms of two-joint muscles are able to walk more efficiently.


Review paper on simulation models, experimental studies and robotic systems with two-joint muscles: Schumacher, C., Sharbafi, M., Seyfarth, A., & Rode, C. (2020). Biarticular muscles in light of template models, ex-periments and robotics: a review. Journal of the Royal Society Interface, 17(163), 20180413, https://doi.org/10.1098/rsif.2018.0413

Balance perturbation experiment with the AMP (in cooperation with TU Delft) Schumacher, C., Berry, A., Lemus, D., Rode, C., Seyfarth, A., & Vallery, H. (2019). Biarticular muscles are most responsive to upper-body pitch perturbations in human standing. Scientific reports, 9(1), 1-14. https://doi.org/10.1038/s41598-019-50995-3

Further current research topics from TU Darmstadt: