Marco Hutter’s ANYmal goes on a mountain hike

Science Robotics magazine has featured a recent paper by Branco Weiss Alumnus Marco Hutter. The article “Quadrupedal robots trot into the wild” shows how deep reinforcement learning enables quadruped robots to traverse challenging natural environments using only proprioception. To this aim, the team brought their four-legged machine called ANYmal to gravel- strewn alpine roads and even more challenging terrain: deformable surfaces such as mud and snow, rubble, thick vegetation, and flowing water – all in view of the startled cows. The artificial animal, ANYmal, made it even on the magazine’s cover.

Read the paper on Science Robotics