Meeting Abstract
The majority of fish use whole-body undulations to power swimming and generate maneuvers. This style of locomotion offers certain benefits including efficiency and stealth, and consequently fish are excellent models for designing autonomous under water vehicles (AUVs). While straight swimming can be generated by simple cyclical motions, unsteady behaviors such as turning are more complex. Previous control strategies for maneuvering in fish robots fall into two major categories. The first adds a lateral offset to the normal locomotor wave, biasing the undulation to the right or the left without interrupting normal swimming. The second imitates the C-start maneuver in fish, in which all body segments deflect simultaneously on one side followed by a rapid, posteriorly propagating straightening, which interrupts typical locomotor body oscillations. We developed a turning model based on the kinematics of routine maneuvers from the Giant Danio (Devario aquepinnatus), which consist of pulses of curvature that start near mid-body and propagate posteriorly. These pulses are non-cyclic events and can be modeled as a transient wave with a speed, amplitude, and width. Using a 3D printed robot, we will be evaluating the performance of the pulse model compared to the offset wave and C-start methods. We have also successfully implemented the pulse, the C-start, and offset control models in a multilink robotic system. Preliminary data shows that the pulse model behaves similarly to the live fish model. All three models are able to execute maneuvers, but further testing will show how the maneuverability, agility, and controllability compare between the turn models.
