⇧ [VIDÉO] You may also like this affiliate content (after the ad)
How do the reflexes of animals allow them to learn to walk in a short time? Researchers from the Max Planck Institute for Intelligent Systems investigated the matter. However, to answer this, they used robotics instead of biology. For the study, they designed a robot that can learn to walk, without instruction, in just one hour.
This is how “Morti” was born, a robot dog designed to study the development of walking in young animals. “ Spinal cord research in living animals is complex. But with a robot, we can model one “Explained Alexander Badri-Spröwitz, who co-authored the publication, in a press release from the Institute. The research was published in the journal machine intelligence nature.
Of course, scientists already know a lot about the motor movement of animals. In fact, as the statement recalls, animals are born with muscle coordination networks embedded in their spinal cord. So they quickly learn to walk to escape from predators. In the beginning, they relied on this relatively basic data, but it allowed them to start moving. Afterwards, they learn to coordinate muscles and tendons more precisely. Their movements gradually become more fluid, but it takes practice.
Scientists claim that these two steps are driven by two mechanisms. In humans and animals, there is something called “ central pattern generators (CPG), ie networks of “neurons” in the spinal cord that produce periodic muscle contractions without brain intervention. This mechanism occurs for repetitive tasks, such as blinking, walking, etc. These are basic tasks, which don’t need to be really connected to the brain. In addition to this, the “reflexes” intervene. These are involuntary movement actions, the scientists explained, which are also hardwired directly into the spinal cord, not the brain.
A virtual spinal cord
Concretely, on a flat surface without traps, the central pattern generator works. Reflexes occur when there is something unusual, a dip, a bump, etc. In the newborn animal, the CPGs are initially insufficient and the animal stumbles, on flat or uneven ground. Over time, he learns how his reflexes and “core pattern” control his movements, and improves. In other words, field experience gradually corrects old gait data. However, knowing the details of this operation is not easy. So the scientists decided to go through robotics.
“ As engineers and roboticists, we seek the answer by building a robot that has reflexes like an animal and learns from its mistakes. », explains Felix Ruppert, one of the authors. So they pair a basic program with a machine learning algorithm. In other words, they started by creating a virtual “spinal cord”, which they attached to the robot dog’s head. The researchers did not give Morti any basic information about the shape of his body, motors, springs…
So the learning algorithm makes it possible to compare the sensory information seen by the robot’s basic program to gradually improve its capacities. The robot dog designed by researchers proved to be able to learn, in the same way as animals, to walk: and faster than an animal, because it only takes him an hour to reach in good coordination of their movements. During this time, the sensor data from the feet (paws) is continuously compared to the landing predicted by the robot’s CPG. When the robot stumbles, the algorithm adjusts the swing speed of the legs, the distance, the duration of contact with the ground, etc.
” We know that these CPGs are present in many animals. We know that reflexes are combined; but how do you combine the two so that animals learn movements with reflexes and CPGs? This is fundamental research at the intersection between robotics and biology. The robotic model allows us to answer questions that biology alone cannot answer “say the scientists.