🤖 Here's what a robot designed by math looks like, without taking inspiration from nature (video)

Most robots are inspired by humans, dogs, or insects. But a team from Duke University has just shown that the most efficient shape might look like nothing familiar.

Their machine, named Argus, has 20 telescopic legs arranged around a central body. This unusual architecture allows it to move in all directions without turning, climb walls, and carry heavy loads. An approach that challenges decades of biomimetic robotics.


Argus has no front or back. Its 20 legs, each equipped with a depth camera at its tip, radiate from a central body. The researchers named their creation after the giant Argus from Greek mythology, who had a hundred eyes. This robot can roll on grass, sand, rocks, and even climb vertical walls. It maintains its balance even after being violently pushed or after losing three of its legs.

The key innovation relies on a mathematical concept called dynamic isotropy. This score, ranging from 0 to 1, measures a robot's ability to accelerate uniformly in all directions. Most current robots, including humanoids and drones, score below 0.6. Argus achieves 0.91, almost the theoretical maximum. This means it can react equally well forward, backward, or sideways.

The researchers achieved this performance by arranging the legs around a particular geometric shape, the regular dodecahedron. This structure with twelve pentagonal faces gives the robot a nearly uniform field of view. The robot does not need to orient itself to move or interact with its environment.

[align=align][/align]

Tests were conducted on the Duke University campus on varied surfaces such as concrete, grass, mud, and sand. Argus cleared obstacles and pushed a cube about 3.3 feet (1 m) on each side. Even with three damaged legs, it continued to move. The researchers also demonstrated its ability to climb between two vertical walls.

This research does not just propose an innovative robot, but a new design method. Instead of copying nature, engineers can now start from deep mathematical principles to create machines with unprecedented capabilities. Chen, co-author of the study, explains that robots don't need to imitate humans or dogs to gain agility. Simply seeking perfect dynamic symmetry can transform robotics.

The implications go beyond a simple demonstration. This mathematical framework could be used to compare different robot bodies and design new ones from scratch. Xia, a postdoctoral fellow in the lab, notes that this approach produces a robot capable of navigating rough terrain, cluttered environments, and even low gravity. It changes what is possible in robotics.