Dynamic shape control in highly inertial regimes

Develop dynamic shape control for multi-segment soft robotic arms operating in highly inertial regimes beyond the capabilities of the current hardware, where increased excitation causes oscillation and reduced convergence stability.

Background

The paper presents a Koopman-based dense model predictive controller for real-time shape control of multi-segment soft robotic arms, with experiments conducted using external motion capture and actuators subject to update-rate and capacity constraints. The reported system achieves dynamic tracking, but the authors note that more aggressive excitation produces increased oscillation and reduced convergence stability.

The unresolved issue concerns extending reliable dynamic shape control to highly inertial operating regimes that exceed the demonstrated hardware capabilities. This limitation is distinct from the paper’s demonstrated free-deformation tracking and moderate-disturbance robustness results, and motivates future improvements in sensing, actuation, and control.

References

This study has the following limitations. First, all wrist stiffness measurements were conducted under quasi-static loading conditions; dynamic behavior during manipulation tasks---such as rapid grasp adjustments or hammer swings---was not evaluated, and the morphological contributions identified here may behave differently under inertial or impact loading.

— Anthropomimetic Soft Robotic Forearm with Independently Articulated Carpal Bones Enabling Human-Like Adaptive Stiffness Modulability  (2609.29176 - Obata et al., 24 Sep 2026) in Limitations of the Present Study; Conclusion

As a result, dynamic shape control for multi-segment soft robotic arms in the highly-inertial regime remains an open challenge for the current hardware.

— Real-Time Shape Control of Multi-Segment Soft Robotic Arms Using Koopman Operators with Global and Local Observables  (2609.03175 - Wang et al., 2 Sep 2026) in Section 6, subsection “Limitations”