How snake-like sea robots mimic nature?

Imagine a robot that moves like an eel, undulating its body to swim in water. This isn't science fiction: researchers are developing continuous robots inspired by anguilliform locomotion, a highly efficient swimming method used by eels. The robot has six segments actuated by pneumatic artificial muscles, similar to those that move real animals. The challenge? Determining the direction of a body that undulates is complex. To solve this, researchers use techniques like dynamic mode decomposition (DMD) and complex-variable delay embedding (CDE). These help predict where the robot will go. To change direction, they test a hypothesis: if the left and right sides are actuated asymmetrically, slightly different fluid velocities are created, altering the movement. Using simulations in COMSOL Multiphysics, they showed the robot can maintain straight-line swimming with a settling time of about 60 seconds. This could revolutionize autonomous marine robots for ocean exploration.

Anguilliform locomotion represents one of the most complex forms of movement in biorobotics, characterized by full-body undulation involving dynamic interactions between fluids and structures. The developed model is based on a continuous body composed of six segments, each actuated by a pneumatic artificial muscle (PAM). These PAMs provide flexibility and force, faithfully replicating the muscle contractions of real animals. Motion prediction is achieved through a technique called dynamic mode decomposition (DMD), which extracts dominant modes from the system's behavior. To enhance prediction, a methodology called complex-variable delay embedding (CDE) was integrated, better capturing nonlinear temporal dynamics. Asymmetric actuation between the left and right sides of the robot generates pressure differences in the surrounding fluid, enabling directional control. Simulations were conducted using COMSOL Multiphysics 6.2, an advanced software for multiphysics modeling. Results show that the CDE-DMD model can accurately predict the robot's trajectory in various scenarios. Furthermore, integrating the predictive model with the asymmetric actuation strategy enabled the development of a feedback control (closed-loop) system that maintains directional stability. In tests with non-zero initial position (step response), the robot successfully returned to straight-line swimming with a settling time of approximately 60 seconds. This type of control could be applied to future marine robots for environmental monitoring missions, underwater infrastructure inspection, or deep ocean exploration.