This post is also available in:
עברית (Hebrew)
Coordinating large groups of robots usually requires complex software, wireless communication, and centralized control systems. But these approaches can become fragile in unpredictable environments where communication links fail, individual units break down, or terrain constantly changes. Engineers have long searched for ways to make robotic swarms more adaptable without relying on a central “brain” to coordinate every movement.
Researchers have now developed a robotic collective that approaches the problem differently, by embedding coordination directly into the robots’ physical behavior. The system consists of dozens of slim robotic modules that move and reorganize themselves through simple mechanical interactions rather than explicit command structures.
Each module measures about 20 centimeters long and contains a small motor that repeatedly changes the robot’s shape between a straight “I” configuration and a curved “U” shape. According to Interesting Engineering, this constant shape-switching generates motion across surfaces. At both ends of the robots are weak Velcro-based connection points that allow nearby modules to attach and detach dynamically while moving.
Individually, the robots move slowly and struggle with uneven terrain. But once linked together, their behavior changes significantly. Chains of connected modules can navigate slopes and obstacles more effectively than isolated units. Instead of following preprogrammed formations, the robots continuously reorganize themselves through physical contact, creating collective movement patterns that resemble flowing materials more than traditional machines.
The researchers describe this as “mechanical intelligence”, where coordination emerges naturally from the robots’ shape and interactions rather than from advanced computation. In obstacle-heavy environments, the swarm-maintained cohesion while still allowing individual connections to break when necessary to avoid becoming trapped.
The system is also designed for resilience. Because there is no central controller, the collective continues functioning even when some robots fail or lose power. Researchers added a simple recovery mechanism in which isolated robots emit an audible buzzing signal that causes nearby units to slow down, allowing disconnected modules to rejoin the group.
From a defense and security perspective, decentralized robotic swarms could eventually support operations in unstable or communication-denied environments where traditional coordination methods are unreliable. Systems that continue functioning despite unit failures or disrupted networks may offer advantages for reconnaissance, search operations, or movement across difficult terrain.
The project reflects a broader shift toward robotics inspired not only by biology, but also by the physics of collective behavior found in natural materials and living systems.
The research was published here.
