Researchers developed fish-inspired robots that maintain swimming patterns in different sizes, ranging from 60 centimeters to 2.9 meters.
Building a larger underwater robot usually requires much more than just scaling up all its parts. Researchers at New York University in the United States have developed an alternative that can preserve swimming movement in machines of significantly different sizes. Inspired by fish, they created three robots measuring 0.6, 1.1, and 2.9 meters that utilize virtually the same mechanical architecture.
The solution aims to address a limitation of existing fish-like robots. Many are designed for a specific function and scale, forcing engineers to rebuild much of the system when they need to develop a larger or smaller machine. In the new concept, only one structural component of the tail needs to be resized to maintain the same swimming behavior.
The study was published on July 3, 2026.
-
5,000-Year-Old Mega-Structure with Basalt Rings and Over 40,000 Tons of Stones Intrigues Archaeologists
-
Scientists Transform PET Bottles into Ingredients for Protein-Rich Cookies
-
Indian Ocean Seafloor Tears Apart for First Time in Front of Scientists: Crust Shifts 4.2 Meters in Just 6 Days, Spilling 160 Million Cubic Meters of Lava and Revealing the Birth of a New Portion of the Planet Live
-
Three Hurricanes Simultaneously Move Across the Pacific: Marie Strengthens Near Mexico, Spanning About 630 km Wide, Karina Reaches Category 4, and Lowell Hits the Maximum Category 5 in a Rare El Niño-Driven Scenario
Tail thickness changes as the robot grows
The project developed by Nana Obayashi and her colleagues drew inspiration from the swimming patterns of cod and mackerel. In these animals, a significant portion of the body curves toward the tail during movement, a pattern found in nature across fish of various sizes.
The robots feature a rigid front section and a flexible tail constructed with fiberglass rods. To produce the necessary undulation, a system of crossed tendons is activated by the motor, forcing the rear section to form a curve resembling an “S.”
As the size of the machine increases, the principle remains unchanged. The main adjustment occurs in the diameter of the tail rods, which become proportionately thicker. This way, the structure maintains flexibility suitable for the new dimensions. The tendon system maintains the same configuration, ensuring that each scale does not require an entirely new design.

Three sizes tested in different environments
The team built prototypes measuring 60 centimeters, 1.1 meters, and 2.9 meters long. Tests were conducted in shallow, narrow, and deep environments to observe how each machine would respond to different conditions.
The smallest robot produced water movement patterns similar to those observed in real fish. More importantly for the project’s goal, all three models exhibited comparable movements when results were adjusted to their respective body sizes.
The difference between the smallest and largest robots reaches nearly five times in length. Nonetheless, the way they moved their tails remained proportionally consistent. This finding provides researchers with a platform to investigate a challenging question often overlooked in traditional experiments: how swimming changes as body size increases.
Larger Robot Maintained Movement but Consumed More Energy
The success in scaling swimming did not entirely replicate when the team analyzed energy efficiency. The two smaller models showed similar results in this regard. However, the 2.9-meter model was less efficient and required a different, more powerful motor to achieve movement.
Researchers believe that factors such as drag and inertia may help explain the increased energy demand, although the cause is not yet fully established.
The experiment demonstrated an important difference: it is possible to preserve the mechanics of swimming while scaling up the robots, but the energy needed to put larger structures into motion does not necessarily follow the same proportionality.
Size Also Affects Agility and Stability
Another contrast emerged when researchers caused deviations in the robots’ trajectories. The smaller model exhibited greater agility, able to alter its movement more easily. However, after being disturbed, it took longer to regain its original orientation.
For the larger models, the opposite occurred. They did not respond as quickly to changes in direction but exhibited greater stability after experiencing a deviation.

This difference suggests that the choice of optimal size may depend directly on the environment and task. A smaller machine may be more suitable for rapid maneuvers, whereas a larger structure may offer advantages when stability is more critical to performance.
Fish-Inspired Robots Avoid Limitations of Propellers
The attempt to replicate aquatic animals is not solely for aesthetic reasons. In certain studies, traditional vehicles powered by propellers may face significant limitations.
Propellers can get tangled in vegetation, disturb sediment on the bottom, and provoke reactions in wildlife. Systems that mimic the propulsion found in fish offer an alternative for situations where it is essential to minimize environmental disturbance.
Obayashi summarizes the proposition based on the diversity of the studied locations: if the environments do not have standardized dimensions and conditions, the tools used to investigate them should not be limited to a single size.
Technique Could Be Applied to Flexible Machines Out of Water
The next challenge is to determine whether energy efficiency can be adapted as easily as achieved in the tail movement.
Even without this answer, the team believes that the principle used in the project could have applications beyond aquatic robots. The strategy involves identifying a decisive structural parameter and altering only that characteristic as the machine’s scale changes.
In the fish-inspired experiment, this element was the thickness of the rods that control the tail’s flexibility. This approach allowed for the construction of machines with significant length differences without redesigning the entire mechanism responsible for swimming.
The result paves the way for a generation of flexible robots that are less dependent on custom designs for each dimension, although the energy performance of larger versions remains a question to be addressed.
