Robot snake slithers brilliantly after much trial and error

STUDENT PROJECT - Intelligent robots - for search and rescue, exploration and surveillance - are a new and challenging field. Unique and versatile, could robot snakes advance the abilities of machines in these fields?

Serpentine robots are a ground-breaking innovation in robotics. These flexible and slithering machines excel at navigating complex and confined spaces, offering immense potential in various fields.

In search and rescue scenarios, for example, robotic snakes are excellent at manoeuvring through narrow and cluttered environments such as collapsed buildings or debris. With their flexible bodies, they can easily navigate tight spaces, climb obstacles and explore inaccessible areas. Equipped with cameras and sensors, these robots transmit real-time data, aiding rescuers in locating survivors and assessing surrounding conditions.

With the objective to undertake a project that would have practical relevance in the real world, this year, a team of EPFL bachelor’s students in the School of Computer and Communication Sciences (IC) undertook a project to make Flipper the robot snake as part of the Making Intelligent Things course.

Flipper is a 1.2-meter-long robot snake crafted from 13 intricately designed 3D-printed pieces. It navigates and explores its surroundings using three distinct modes of motion: Inchworm, Concertina, and Undulation. The Inchworm mode enables precise inch-by-inch movement, the Concertina mode mimics the graceful accordion-like motion and in Undulation mode, Flipper effortlessly slithers with a sinusoidal movement.

"Our project had to be composed of a balanced amount of software and hardware as well as being both complex yet feasible in the eight-week time frame that we had," explained Leila Sidjanski, a second year Bachelor in Communication Systems student and one of the team’s members. "It was overwhelming at first but we did find a snake project online that inspired us and we had some amazing guidance from a PhD robotics student here at EPFL."

The team had a solid understanding of algorithms, programming languages and system design, however no direct experience in building robotic systems. They embraced the opportunity to expand their expertise.

"The first thing we started on was the design of the pieces that we were going to print even though we have no background in 3D design. We also had to figure out how the snake would move smoothly and this was very difficult in the code. We had to continually tweak the parameters and it was a lot of trial and error. Finally, another very practical challenge was allocating and timing the tasks between us, because you can’t start testing the snake if the hardware isn’t finished and you can’t start testing its movement if you haven’t connected the cables together," said Lea Grieder, another member of the team who is studying for a Bachelor in Computer Science.

Professor Christoph Koch teaches the course and explained that making a robot move like a real biological animal is challenging, particularly in the case of snakes whose movement patterns are particularly subtle.

"Real snakes rely on the microstructure of the scales on their bodies to move, something we can’t replicate, so a pattern of movement that looks right might keep the robotic snake wiggling on the spot without any forward movement. The team talked to actual researchers in the field and read scientific publications on the topic and then made some very clever design choices that led to success. They did all this out of their own drive without being instructed to do so."

For the team, one of the most gratifying things about the course and their project was being able to show what mathematics can do!

"Until now the math in our studies has been abstract. For me, math is imaginary, you can’t really see it and through Making Intelligent Things we were able to do something physically, put all of our knowledge to the test and build something was really cool. We can show what math can do! Our families were amazed at what we built in such a short amount of time and saw that we were enjoying ourselves immensely," added Sidjanski.

The team hopes to make some further improvements and additions to Flipper and continue showcasing it to wider audiences. And would they recommend Making Intelligent Things?

"If I was asked that question in the first two weeks of the course I would have said no because we were thrown in the deep end, it was basically, ’figure it out’! But as you go along you learn a lot - 3D printing, electronics, soldering, how to navigate organization by yourself - and it’s certainly one of the courses in which we have learnt the most in such a short time. It was so enjoyable and we wanted to spend most of our time working on it. I think it’s quite amazing and that’s what makes us even prouder of our project," concluded Grieder.

Professor Koch sings the praises of his students and says that the projects they develop in this course continue to surprise him, "In the case of Flipper the robot snake, I didn’t expect it to work as well as it has! Several of the other teams also made really impressive progress in very short amounts of time, and we had quite a few brilliant - and brilliantly working - projects in the end!"

The five members of the Flipper the Robot Snake team were: Robin Bochatay, Cherilyn Christen, Lea Grieder, Giovanni Ranieri and Leila Sidjanski.

The six other projects developed this year by students taking the Making Intelligent Things course were:
* a physical Mario Kart game
* a 2D plotter for disabled people (drawing on facial expression)
* a self-balancing unicycle
* a SLAM (simultaneous localization and mapping) robotic vehicle
* StrumMaster (a self-playing acoustic guitar)
* an automated warehouse with autonomous forklift