Robot fish

I’ve made a few robot rays and robot fish. I originally started doing this while working at Animal Dynamics Limited and so my early innovations belong to them. I hope they use them well. All these new ones have been redeveloped and reimagined from scratch. I have applied for various patents related to these designs. See Patent application: GB2112894.7

The later videos are purposely filtered and kept very short to avoid giving away too many secrets. However, in general, I’m very happy to discuss these things with anybody who is interested – so please contact me if you’d like to discuss. I spent around 6 months using a suite of wing and motor designs to try to home in on the design that might best replicate the type of swimming seen in nature – the philosophy being that this would be a good place to start to find an optimal design. I tried hundreds of tethered thrust tests after a few free swimming exercises. I found the tethered tests where the fish accelerates from a standing start to cruise where the most informative.

The following video is close to the optimal design.

There was a load of scientific measurement of these rapid prototypes, and in summary their function was to act as intermediary steps toward the optimal design.

Graph of thrust against power (current), for many various alterations in software and hardware.

The primary focus of the testing was to develop the structure and function of the wings to replicate a natural ray of a similar size. There are a few very good scientific studies of swimming rays, where the pertinent details are recorded. By using point tracking on the videoed results, I was able to get very accurate measurements of wing dynamics and overall movements. There was also a full IMU on the vehicle and direct reports from the motors of their current draw – this interactive response to thrust is a key aim of all these rays. An example of comparison with natural fish is as follows: