Robotic dog runs (almost) entirely on its own

- EN - DE- FR- IT
© Alain Herzog/EPFL
© Alain Herzog/EPFL

For his Master’s project at EPFL, MickaŽl Achkar compiled data on the movements of dogs to develop a robotic version of the animal that, once set in motion, can run without assistance.

Engineers at EPFL’s Computational Robot Design & Fabrication Lab (CREATE), headed by Prof. Josie Hughes, are coming up with new ways of building robots possessing never-before-seen capabilities. For instance, Hughes and two other researchers used ChatGPT to design a robotic gripper for harvesting tomatoes. And MickaŽl Achkar used motion-capture data from live dogs to build a robotic one. More specifically, Achkar studied dogs’ biological mechanisms to create a smarter robot design and build a prototype that can run by itself once set in motion.

"I wanted to engineer a robot with animal characteristics, bearing in mind that animals - like humans - move in a huge variety of ways," says Achkar. "But most of these movements are executed by just a few joints." He therefore drew inspiration from animals’ motor control processes to guide him in his robot design.

Achkar could have chosen just about any animal - a grasshopper, mouse, elephant or cheetah, for example - but a dog turned out to be the obvious choice. "We found a vast dataset on the motion of dogs, and it was even available in open source!" he says. The first step was to extract data on dogs’ synergistic movements and then structure the data so they could be "summarized" in a meaningful way, through a method known as principal component analysis. This basically entailed grouping the data into several vectors describing the main axes of dog motion, and using this information to establish exact specifications for the robot.

Metal, pulleys, cables and screws

Achkar’s robotic dog has bilateral symmetry. Each of the robot’s four legs has three joints, and each joint is coordinated with the others. This latter feature is the added benefit that enables Achkar’s robot to run in the same way as - and with all the agility of - a real dog. To build the prototype, Achkar used metal rods as the bones, 3D-printed pulleys as the joints, thin cables as the tendons and a few screws to hold it all together.

The engineers bought a treadmill to test their prototype. They discovered that once the robot got going, it could run autonomously without having to activate its control motors. "At first we thought it might’ve been a fluke," says Achkar. "So we changed the design slightly and tested the robot again - and it couldn’t run anymore." However, the research team did end up adding a counterweight, similar to a pendulum, so that the robot could stay in motion once it started. "The counterweight uses resonance to inject energy," says Achkar. Francesco Stella, a PhD candidate at CREATE and the project supervisor, adds: "We designed the robot’s body to be able to respond automatically, much like a trout starts swimming automatically when placed in water."

Joints moving in synergy

The robot’s control motors are nevertheless useful for achieving a broader range of motion. For instance, it can jump and overcome obstacles without the help of its counterweight. "We’d like to push our design further with the motors, but for now the prototype isn’t very robust," says Achkar. That didn’t prevent him from putting the mechanical dog to the test, such as by placing a stick between its legs to see how it would respond. Unfazed, the robot automatically resumed its graceful gallop. And on the treadmill, it easily reaches a speed of 6 km/h.

"Our goal isn’t to compete with ultra-high-tech robotic dogs, but rather to explore bio-inspired robot designs," says Achkar. "This entails honing a robot’s fundamental design and modifying its passive proprieties so that only simple control systems are needed - all while maximizing the robot’s capabilities. What we’ve done here - engineering the joints to work in synergy - has already proven useful for creating robotic hands and other body parts."

Achkar has submitted his research paper to a scientific journal for publication, and it should appear in the coming months. Now that he’s completed his Master’s in robotics engineering, Achkar plans to return to Montreal. He came to EPFL from Canada after getting a Bachelor’s degree in mechanical engineering from McGill University. Why did he choose EPFL? Because it offered an excellent education and was located in a French-speaking part of Europe. It also gave him the chance to discover the exciting world of robotics.