AI-Powered Robots Evolve and Self-Repair After Being Chopped into Pieces
Engineers in the United States have created groundbreaking AI-powered robots that can evolve, reconfigure themselves, and recover from severe damage, such as being chopped into pieces. These alien-like machines, described as 'legged metamachines,' are built from autonomous modules that snap together in various forms, offering unprecedented resilience and adaptability in real-world environments.
Modular Design and AI-Driven Evolution
The research, conducted by engineers at Northwestern University and published in the journal Proceedings of the National Academy of Sciences, focuses on modular hardware. Each module functions as a small robot with a motor, battery, and computer, capable of rolling, turning, and jumping independently. When combined, these modules form larger robotic systems with agile movements, designed by artificial intelligence to generate unusual body shapes not typically considered by human engineers.
Using an evolutionary algorithm that mimics natural selection, the AI simulated and selected the most effective configurations. Over iterations, designs evolved to include legs, spines, or tails, resulting in robots that move in striking ways—undulating like seals, bounding like lizards, or springing like kangaroos. The metamachines can flip themselves upright, hop over obstacles, and perform acrobatic maneuvers like spinning in mid-air.
Resilience and Self-Repair Capabilities
A key innovation is the robots' ability to self-repair. Because each system comprises multiple self-contained modules, damaged parts can separate and continue operating, preventing total failure. In outdoor trials, the robots navigated rough terrain including gravel, grass, sand, and mud, maintaining functionality even after losing limbs. Detached modules could roll back and reconnect, showcasing robust recovery from injuries that would be fatal to conventional robots.
Dr. Sam Kriegman, who led the study, emphasized the significance: 'These are the first robots to set foot outdoors after evolving inside of a computer. They can survive being chopped in half or cut up into many pieces, with every module becoming an individual agent.' This modular approach allows for rapid assembly, repair, and redesign, enabling immediate movement across unstructured environments without additional programming.
Implications for Future Robotics
The research builds on earlier work from Dr. Kriegman's laboratory, where AI designed simple robots from scratch. This new study advances the field by creating robots that sense their surroundings, compute, and learn, pointing toward a new class of adaptable robots for unpredictable settings. Supported by Schmidt Sciences and the National Science Foundation, the project highlights how evolutionary algorithms can reveal innovative designs beyond human imagination, potentially transforming applications in exploration, disaster response, and beyond.
