Inspired by ants, which can collectively create impressive structures without any central control or pre-prepared plans, researchers have successfully transferred a similar approach to robotic systems. Developed by researchers at Harvard University's John A. Paulson School of Engineering and Applied Sciences (SEAS), the robotic ants named "RAnts" offer a self-organizing structure without a central command system or detailed programming.
These robots, just like real ants, can work together to build complex structures and dismantle them when necessary. Moreover, they do so without needing any plan, leader, or direct communication mechanism.
RAnts Project Based on the "Exbodied Intelligence" Concept
The most striking concept behind this system is "exbodied intelligence," which can be translated as "body-external intelligence" or "environmental intelligence." According to this approach, intelligence does not originate solely from the software or hardware within the robots; rather, it emerges from the continuous interaction between the robot and its environment. In other words, the reasoning process here arises not within individual units, but within the entire collective system. According to Professor L. Mahadevan, one of the study's authors, this model clearly demonstrates how simple rules can combine to create complex and harmonious behaviors.
In the real world, ants leave chemical trails called pheromones to communicate. These trails determine which direction other ants should go or where they should work. The Harvard team transferred this mechanism to the digital environment, developing a light-based system called "photormone." The robots move by sensing these light fields and update their behavior according to changes in the environment. This process is based on the principle known in biology as "stigmergy," which enables individuals to communicate indirectly by reacting to environmental changes.
The RAnts system is built upon very simple rules. Robots follow specific signals, carry blocks, and drop these blocks when appropriate conditions are met. However, when these simple rules come together, a highly organized and flexible system emerges. Furthermore, only two parameters need to be adjusted to change the system's behavior: sensitivity to light signals and block dropping/picking up thresholds. When these parameters are changed, the robots can instantly switch roles, moving from construction to demolition.
According to the researchers, this approach could offer a significant alternative for autonomous robotic systems in the future. This is because the model demonstrates that complex tasks can be accomplished without central control, solely through local interactions and simple rules. This study also presents a different perspective on the future of artificial intelligence. Here, intelligence arises from the dynamic interaction between multiple agents and the environment, rather than from the capacity of an individual system. This suggests that larger and more complex "intelligence" systems might emerge in ways far different from what we have considered until now.
0 Comments: