Fig. 8: Batoid-robot size and actuator type influence normalized speed.

a Normalized speed (in CLs⁻¹) versus CL shows differing levels of performance for different actuator types. These include tissue engineered (green), IPMC (pink), and SMA (yellow), with slower speeds at small scales, while DEA (dark pink) and HASEL (red) performed well at small scales. Tensegrity (blue) performed well at larger scales, pneumatic (orange) performed better at smaller scales, and servo (white) performed well primarily at larger scales. Biological rays (black) performed better than all ray-inspired robots. There is a gap in performance between CL = 15 to 35 cm, as seen in the grey section. There is also an upper bound (black dashed boundary) on robot performance because of servo limitations. Characteristic length (CL) is defined as \(\frac{{\rm{Width}}+{\rm{Length}}}{2}\) to account for discrepancies in aspect ratio between robots. b Normalized speed (CLs⁻¹) versus CL for different servo actuator mechanisms, including a single-servo along the leading edge (SSLE), multiple aligned servos (MAS), and multiple servos with complex bending mechanisms (MSCB). Less complex single-servo robots show highest maximum speeds, indicating a tradeoff between complexity and performance. No drivetrain servo-driven robots reported sufficient data for inclusion in this plot^{17,19,24,40,41,42,44,45,46,58,59,61,62,63,68,70,71,73,74,75,76,77,79,80,82,84,85,86,87,120,121,122,123,124}.