Robot-ants that communicate and work together


Monday, 15 July, 2019

Robot-ants that communicate and work together

A team of researchers at École polytechnique fédérale de Lausanne (EPFL) has developed tiny robots that are inspired by ants: they can communicate with each other, assign roles among themselves and complete complex tasks together. The researchers have just published their work in Nature1.

Individually, ants have only so much strength and intelligence. However, as a colony, they can use complex strategies for achieving sophisticated tasks to survive their larger predators.

At EPFL, researchers have reproduced this phenomenon, developing tiny robots that display minimal physical intelligence on an individual level but that are able to communicate and act collectively. Despite being simple in design and weighing only 10 grams, each robot has multiple locomotion modes to navigate any type of surface. Collectively, they can quickly detect and overcome obstacles, pass them and move objects much larger and heavier than themselves.

These three-legged, T-shaped origami robots are called Tribots. They can be assembled in only a few minutes by folding a stack of thin, multi-material sheets, making them suitable for mass production. Completely autonomous and untethered, Tribots are equipped with infrared and proximity sensors for detection and communication purposes. They could accommodate even more sensors depending on the application.

“Their movements are modelled on those of Odontomachus ants,” said Zhenishbek Zhakypov, the first author. “These insects normally crawl, but to escape a predator, they snap their powerful jaws together to jump from leaf to leaf.”

The Tribots replicate this catapult mechanism through an elegant origami robot design that combines multiple shape-memory alloy actuators. As a result, a single robot can produce three distinctive locomotive motions — crawling, rolling and jumping both vertically and horizontally — just like these creatively resilient ants.

Despite having the same anatomy, each robot is assigned a specific role depending on the situation. ‘Explorers’ detect physical obstacles in their path, such as objects, valleys and mountains. After detecting an obstacle, they inform the rest of the group. Then, the ‘leader’ gives the instructions. The ‘workers’, meanwhile, pool their strength to move objects.

“Each Tribot, just like Odontomachus ants, can have different roles,” said Professor Jamie Paik. “However, they can also take on new roles instantaneously when faced with a new mission or an unknown environment, or even when other members get lost. This goes beyond what the real ants can do.”

In practical situations, such as in an emergency search mission, Tribots could be deployed en masse. And thanks to their multi-locomotive and multi-agent communication capabilities, they could locate a target quickly over a large surface without relying on GPS or visual feedback.

“Since they can be manufactured and deployed in large numbers, having some ‘casualties’ would not affect the success of the mission,” Paik added. “With their unique collective intelligence, our tiny robots are better equipped to adapt to unknown environments. Therefore, for certain missions, they would outperform larger, more powerful robots.”

The development of robots for search-and-rescue applications and the study of collective robotics are key research areas within the NCCR Robotics consortium, of which Jamie Paik’s lab is part.

Reference
  1. Zhakypov Z, Mori K, Hosoda K and Paik J 2019, Designing minimal and scalable insect-inspired multi-locomotion millirobots, Nature
Related News

Emerson offers solution to reduce energy costs and emissions

Energy Manager is designed to simplify electricity monitoring, tracking real-time use to identify...

New robotics and automation precinct opens in WA

The WA Government has officially opened what it says will be Australia's largest robotics and...

International robot federated learning project a success

The FLAIROP international research project has shown AI federated learning across multiple...


  • All content Copyright © 2024 Westwick-Farrow Pty Ltd