Organizations: C. Gay, P. Trunin, D. Cafiso and L. Beccai are with the Soft BioRobotics and Perception Lab of the Istituto Italiano di Tecnologia, Via Morego 30 16163, Genova, Italy · Y. Xu and M. Taghavi are with the Department of Bioengineering, Imperial College London, SW7 2BX London, U.K.
Electro-Ribbon Actuators (ERAs) are lightweight flexural actuators that exhibit ultrahigh displacement and fast movement. However, their embedded sensing relies on capacitive sensors with limited precision, which hinders accurate control. We introduce OS-ERA, an optically sensorized ERA that yields reliable proprioceptive information, and we focus on the design and integration of a sensing solution without affecting actuation. To analyse the complex curvature of an ERA in motion, we design and embed two soft optical waveguide sensors. A classifier is trained to map the sensing signals in order to distinguish eight bending states. We validate our model on six held-out trials and compare it against signals' trajectories learned from training runs. Across all tests, the sensing output signals follow the training manifold, and the predicted sequence mirrors real performance and confirms repeatability. Despite deliberate train-test mismatches in actuation speed, the signal trajectories preserve their shape, and classification remains consistently accurate, demonstrating practical voltage- and speed-invariance. As a result, OS-ERA classifies bending states with high fidelity; it is fast and repeatable, solving a longstanding bottleneck of the ERA, enabling steps toward closed-loop control.
Figures & tables
Fig. 1: A. Optically Sensorized Electro-ribbon actuator during movement. B. Working principle of the bending optical sensor C. Working principle of ERA.
Fig. 2: A. Curvature of the two areas to sensorize of the upper electro-ribbon during the loading (20 g mass for 51.75 mm with a voltage of 8 kV). B. Design of the upper ribbon surface equipped with the soft optical sensors, S1 and S2. Two emitters are positioned in the clips near the anchoring point, while two receivers are placed in the distal clips.
Fig. 3: A. Fabrication process B. OS-ERA sample configuration C. Example of encapsulated component D. Insight image of the waveguide superficial structure
Fig. 4: A. Mean and standard deviation of the three training trials (12.3 g, 4 kV) for the two sensors (in black and in red). The colored dots are equally spaced in time (except for the last one) and represent the eight selected states during actuator contraction. The green band indicates the time interval during which the sensor buckles. B. Output of the SVM model showing the two sensor signals on the axes, the colored areas and decision boundaries for the eight classes, and the corresponding training samples (dots).
Fig. 5: A-H. Prediction example during OS-ERA movement (12.3 g, 5 kV). The bottom-left number indicates the predicted state, while the colored ribbon above the actuator represents the corresponding curve (saved from training video). The images show frames extracted from the video (12.3 g, 5 kV) at the moments of each prediction. I. Signals’ trajectories of all testing trials on the SVM model map. The three blue trajectories correspond to the testing trials performed at 3 kV, while the three red trajectories correspond to those performed at 5 kV. The colored dots represent the training data obtained from the three trials performed at 4 kV.
School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China · State Key Laboratory of Mechanical System and Vibration, School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China