Robot dog is used to understand how animals learn to walk

A baby giraffe, like other quadrupedal wild animals, must learn to walk on its legs as quickly as possible to avoid predators. However, getting to grips with the precise coordination of leg muscles and tendons takes some time.

Baby giraffe learning to balance on its legs and to walk. Image: Mary Ann McDonald – Shutterstock

They are born with muscular coordination networks endowed with rigid wires located in their spinal cord on which they depend to stand and move through neural reflexes.


Although a little more basic, motor control reflexes help the puppy avoid falling and injuring himself during his first few attempts at walking. Then, more advanced and precise muscle control is practiced, until eventually the nervous system is well adapted to the muscles and tendons of the animal’s legs so that it can keep up with the adults.

Researchers at the Max Planck Institute for Intelligent Systems (MPI-IS) in Stuttgart, Germany, conducted a study (published in the journal Nature Machine Intelligence) to discover how animals learn to walk and overcome stumbling blocks. For that, they built a four-legged robot, the size of a medium-sized dog, which helped them figure out the details.

Researchers in Germany have developed a robot dog to understand the process of learning to move in quadrupedal puppies. Image: Max Planck of Intelligent Systems (MPI-IS)

“As engineers and robotics professionals, we seek the answer by building a robot that reflects like an animal and learns from mistakes,” Felix Ruppert, former PhD student in the dynamic locomotion research group at MPI-IS, said in a statement. “If an animal stumbles, is that a mistake? Not if it happens once. But if it stumbles often, it gives us a measure of how well the robot walks.”

Learning algorithm optimizes robot dog virtual spinal cord

A Bayesian optimization algorithm drives machine learning: the measured information from the foot sensor is combined with target data from the modeled virtual spinal cord running like a program on the robot’s computer. It learns to walk by continuously comparing information sent and received from the sensor, running loops reflexes and adapting their motor control patterns.

The learning algorithm adapts control parameters of a Central Pattern Generator (CPG). In humans and animals, these central pattern generators are networks of neurons in the spinal cord that produce periodic muscle contractions without input from the brain.

Central pattern generator networks assist in generating rhythmic tasks such as walking, blinking, or digestion. Furthermore, reflexes are involuntary motor control actions triggered by encoded neural pathways that connect sensors in the leg with the spinal cord.

As the young animal walks on a perfectly flat surface, CPGs may be sufficient to control spinal cord movement signals. A small bump on the ground, however, changes the walk. Reflexes kick in and adjust movement patterns to keep the animal from falling.

These momentary changes in motion signals are reversible, or “elastic,” and motion patterns return to their original configuration after the disturbance. But if the animal does not stop stumbling through many cycles of movement – ​​despite active reflexes – then the movement patterns must be repaired and made “plastic”, that is, irreversible.

Morti, the robot dog created to understand the locomotion learning process of quadrupedal animal puppies. Image: Max Planck of Intelligent Systems (MPI-IS)

In the newborn animal, the CPGs are initially still not adjusted enough and the animal stumbles around, both on even and uneven terrain. The animal, however, quickly learns how its CPGs and reflexes control the muscles and tendons of the legs.

The same goes for the robot dog called “Morti”, which, in addition, optimizes its movement patterns faster than an animal, in about an hour. Your CPG is simulated on a small, lightweight computer that controls the movement of your legs.

The virtual spinal cord is placed on the quadruped robot where the head would be. During the hour it takes for the robot to walk smoothly, its foot sensor data is continually compared with the expected stumble predicted by the CPG.

If that happens, the learning algorithm changes how far and how fast the legs swing back and forth and measures how long he stays on the ground. Adjusted movement also affects how well the robot can utilize its compatible leg mechanics. During the learning process, the CPG sends adapted motorized signals so that the robot starts to stumble less and optimizes its walking.

Read more:

In this structure, the virtual spinal cord has no explicit knowledge about the design of the robot’s legs, its motors and springs. Knowing nothing about the physics of the machine, it lacks a “model” robot.

“Our robot is practically ‘born’ knowing nothing about its autonomous legs or how they work,” explains Ruppert. “The CPG resembles a built-in automatic walking intelligence that nature provides and that we transfer to the robot. The computer produces signals that control the leg motors, and the robot initially walks and stumbles. Data flows back from the sensors to the virtual spinal cord, where sensor and CPG data are compared. If the sensor data does not match the expected data, the learning algorithm changes the walking behavior until the robot walks well without stumbling. Changing the CPG output, keeping reflexes active and monitoring robot stumbling, is a central part of the learning process.”

Low power consumption makes mechanism more viable

While quadruped industrial robots from prominent manufacturers, which have learned to run with the help of complex controllers, are very energy hungry, Morti’s computer needs only five watts to walk.

“We cannot easily search the spinal cord of a living animal. But we can model one on the robot,” says Alexander Badri-Spröwitz, co-author of the study and head of the Dynamic Locomotion Research Group, in Max Planck. “We know that these CPGs exist in many animals. And we know that reflexes are built in, but how can we combine both so that animals learn movements with reflexes and CPGs? This is fundamental research at the intersection of robotics and biology. The robotic model gives us answers to questions that biology alone cannot answer.”

Have you watched our new videos on YouTube? Subscribe to our channel!