In this short report we have presented the first step towards the design and construction of a novel actuator for small swimming machines. Though we have used rough models, the results show that we are pointing in the right direction. The design shown here is not necessarily the best in reducing the actuation needed. For example, one could think of using the instability of the center to initiate motion, by forcing the system through its bifurcation. This could be achieved by on-line modification of the distance between the supports or by bringing the permanent magnets closer. Additionally, placing a permanent magnet perpendicular to the plane at the origin could be used to further reduce the frequency of the orbits or to control the offset in a more sensitive way than the one shown here. Our model includes dissipation proportional to the velocity and therefore the role of dissipation is marginal. More detailed models of the fluid dynamics and the bending of the fin will surely bring dissipation into a more primary role in the behavior of the system. In addition, thrust, heat dissipation and energy consumption could be estimated in such multi-physics models. We have shown how a simple PID controller could perform reasonably when information about the phase portrait of the system, like the dependence of frequency with amplitude, is included. The use of adaptable frequency oscillators or standard model-based controllers (like feedback linearization), could improve performance and reduce these requirements. Additionally, a system that is too flexible does not possess orbits without a controller. Such a controller would requires large amount of actuation, since it is forcing the system to behave unnaturally. Therefore existence of orbits can be exploited to reduce energy consumption. This stresses the fact that passive dynamics are a key to improve the way we control and design our robots. Controlling the force between solenoids and moving magnets, brings several challenges on the design of the electrical circuits due to the changes in impedance. Another interesting aspect of the problem that will be addressed in further studies. We understand that results obtained solely from simulations are as 'words without actions, however the use of simple models can help us evaluate the feasibility of certain designs. In our particular case, a device with low friction could be driven with a tuned PID, a frequency lookup table and a solenoid with 800 turns, consuming about 20mA.
Magneto-mechanical actuation model for fin-based locomotion
2010
13 Seiten, 6 Bilder, 1 Tabelle, 21 Quellen
Conference paper
English
BASE | 2011
|Fast sensing and adaptive actuation for robust legged locomotion
DataCite | 2023
|Tandem actuation of legged locomotion and grasping manipulation in soft robots using magnetic fields
BASE | 2020
|