Augmented Human Balance Simulation

This video illustrates the potential of supernumerary robotic limbs (SLs) in improving the user's balance through 2 simulation results (simulations performed in Matlab R2024a © ). In these simulations, SLs were controlled to minimize the left-right displacement of the center of mass (CoM) of the human-SLs system, while the simulated human trunk was subjected to a sinusoidal (left-right) torque input at the lumbar spine level. The SLs cost function used to obtain these results is as in Equation 1 of Verdel et al (2024) (see below).

*7--13 seconds, Vertical solution:* This is a first example of equilibrium that can be obtained with SLs to prevent the loss of balance of the user. This solution is costly in terms of motor torques from the SLs, which implies high human-SLs interaction efforts. Furthermore, implies an upward displacement of the human CoM that could results in an overall reduced balance stability.

*14--24 seconds, Horizontal solution:* In this second example, the SLs oscillate around a horizontal posture, which is reminiscent of humans trying to maintain balance in an unstable situation (e.g. walking on a rope). This solution is slightly less efficient than the vertical one in minimizing the displacement of the CoM. However, it also requires less torque inputs from the SLs and leads to a smaller upward displacement of the human CoM, implying a smaller reduction of the balace stability.


_Verdel, D., Eden, J., Cervantes Culebro, H., Pinardi, M., Di Pino, G., Souères, P., Mehring, C., Burdet, E. (2024). A predictive coding framework for safe and versatile control of supernumerary robotic limbs. Arxiv. doi: TBD_

Скачать видео