Germany
Ahsan Tanveer
This project investigates a fish-inspired approach to proprioception in soft underwater robots, asking whether a soft robotic fish can estimate its own body shape and motion using only signals from its actuation system rather than additional embedded sensors. In biological fish, body motion is not sensed only through external perception; it is also informed by the coupling between muscle activation, body bending, and interaction with the surrounding water. This project adapts that principle to a robotic fish by using intrinsic actuation signals, such as motor current, motor position, input voltage, and actuator command history, to infer how the flexible body bends during swimming. The underlying idea is that changes in tail-beat frequency, actuator effort, swimming speed, and water resistance alter these signals in ways that can be related to body deformation. Experiments will be conducted on a soft robotic fish while recording external motion-tracking data as ground truth. By testing different tail-beat patterns, swimming speeds, and flow conditions, the study will determine when actuation signals provide reliable estimates of body motion and when they become insufficient. The goal is to develop a reproducible method for evaluating bio-inspired sensorless proprioception in soft underwater robots and to clarify how actuation signals can be used to estimate motion in fluid environments.
Robotics; Biology; Movement; motion
Ahsan Tanveer is a doctoral student at Peking University, where he works in the Intelligent Biomimetic Design Lab on soft and underwater robotics. During his ZUKOnnect Fellowship at the University of Konstanz, he is developing a sensorless proprioception method for soft robotic fish that estimates body motion from actuation signals.