DEOP: By the Numbers:
Total Weight: 31 kg
Approximate Footprint: 1.1m x 1.2m x 1.3m
Hard Cost: $1,286.12
Rig Frame
The final design incorporates anodized 80/20 brand anodized aluminum extrusions for the base of the structure, selected for their ease of construction and modularity. The rig gets bolted into to the deck with 80/20 brand brackets.
"Brain" of the Rig
The rig runs off of a Raspberry Pi microprocessor, which is housed within a Pelican box and connects to the sensors onboard the rig.
Biteplate
The biteplate is constructed of machined high density polyethylene (HDPE), selected for strength and manufacturability. It incorporates a spring-loaded design, with stainless steel compression springs nested in pockets machined in the biteplate. The outsides of the biteplate are covered in industrial strength neoprene, which can be replaced after repeated use/wear. The biteplate uses a marine-grade Hall effect sensor to trigger a signal to initiate a test. The Raspberry Pi then communicates to the hydrophone to play a tone for the dolphin.
Paddles
One of the paddles is then selected and pressed by the dolphin, which triggers a signal via an accelerometer at the pivot point which registers when the angle of the paddle changes. The rig mounts to the dock with brackets and screws drilled into the dock.
Performance Summary
During the training session, a dolphin was first asked by the trainer to bite on the biteplate while it was held by the sponsor (not attached to the rig). The dolphin performed the behavior with no issues, and compression of the biteplate was observed multiple times. The biteplate was then attached to the rest of the rig and held at the surface. The dolphin bit the biteplate when asked to do so. The rig was then fully submerged, and the dolphin was asked to bite on the plate. At this point, the dolphin had a somewhat hard time biting on the biteplate (he appeared to be confused about what was being asked of him). The trainer working with the dolphin said that it is likely that fully implementing the rig will require a full day of training for the dolphin to get used to the new setup, since the rig has the biteplate located slightly under the deck, while the original did not. (Having the biteplate under the deck was motivated by the angled arms that hold the biteplate needing to accommodate the dolphin’s echolocation anatomy). This training would include lowering the biteplate in intermediate increments until it is at the desired spot. When asked to hit the paddles, the dolphin did so with no issues. The paddles moved as expected, deflecting roughly 10 degrees upon press (5 degrees is the threshold set in the code to register a paddle press), and were restored to vertical well within a second of being released. The rig mounted to the deck with no problems. It sat flush against the dock and was bolted firmly in place. The trials were video recorded with a submerged GoPro, and upon watching the video, the sponsor confirmed that the dolphin’s reaction to the rig was positive and promising.
An above-water electronics demo was also done for the sponsor. The Raspberry Pi was connected to a monitor, and when the biteplate was pressed, a feed of the speakers playing a tone and accelerometer readings being taken was shown on the screen. During training, an observation was made that the dolphins sometimes “chomp” (release and re-bite) on the biteplate. As such, a recommendation was made to turn off Hall effect sensor readings during the response window for each individual trial, which was implemented and confirmed functional after final testing.
The Dolphin Enrichment and Operant Panel successfully meets or exceeds all initial performance requirements, and as such will be a very useful tool for Sean Coffinger and the rest of the researchers at the SPAWAR Point Loma dolphin facility. The feedback received from dolphin trainers and researchers at the facility affirmed that the system has the potential to revolutionize dolphin research and training at the base. There are currently plans to produce more Dolphin Enrichment and Operant Panels as well as add-ons to the existing one. Full implementation of the rig is likely to occur within the next few weeks, as all of the dolphins are trained to use the rig.