Original title: Real-Time Shape Control of Multi-Segment Soft Robotic Arms Using Koopman Operators with Global and Local Observables
Jiahe Wang, Eron Ristich, Sultan Haidar Ali, Eric weissman +4 · September 2026
Soft robotic arms bend continuously to move safely around obstacles. In tight spaces, guiding just the arm's tip is not enough; you must control the shape of the entire body. However, steering every segment is very difficult because soft arms bend in complex, nonlinear ways. Gravity, inertia, and pulling between connected parts make control even harder as more segments are added. Older methods look only at overall shape errors across the whole room. To fix this, researchers built a new real-time controller. Their system tracks both the robot's overall shape and the local bending of individual segments. In computer tests, the method handled arms with up to 10 independently actuated segments. In real-world physical tests, it guided 3- and 5-segment arms reaching tip speeds up to 0.6 m/s. Without retraining, it handled distal payloads up to 400 g, recovered from a 7 N sideways disturbance, and succeeded in a confined-space demonstration.
When you try to shape a piece of wet clay into a curve, pressing one spot changes the rest. The difficulty is that the clay deforms unpredictably, making it hard to get the exact shape without trial and error. In this paper, the soft robot arm behaves like that clay, and the controller uses both overall and segment-specific information to bend it accurately in real time.
This enables soft robots to safely interact in confined spaces like disaster sites or during surgeries. People using these robots can perform tasks requiring bending around obstacles without damage.
The abstract doesn't specify limitations, but it mentions evaluation only with numerical and physical experiments on up to 10 segments and tip speeds up to 0.6 m/s. It doesn't discuss performance in other conditions or with different robot designs.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Real-Time Shape Control of Multi-Segment Soft Robotic Arms Using Koopman Operators with Global and Local Observables — Jiahe Wang, Eron Ristich, Sultan Haidar Ali et al.
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.