Four-Legged Robot Learns to Navigate Rough Terrain
Walking on uneven ground can be a challenge for the best of us, let alone if you are a four-legged robot! Researchers from the Dynamic Legged Systems lab, have released a video showing their robot going through some rigorous testing on uneven ground.
The Dynamic Legged Systems (DLS) lab is based in the Istituto Italiano di Tecnologia (IIT). The DLS lab specializes in the design and control of legged robots for unstructured environments.
The robotics engineers have a focus on robots with high performance, speed, and precision. The lab has produced a paper detailing their work on visual feedback when traversing rough terrain.
Visual feedback is a vital asset for robots to have to develop terrain awareness. Hoover robot locomotion on rough or changing terrain is very difficult.
Their latest work presents a real-time, dynamic foothold adaptation strategy based on visual feedback. The lab's method involves training the robot to adjust the landing position of its feet in a fully reactive manner, using only onboard computers and sensors.
According to the paper ‘the correction is computed and executed continuously along the swing phase trajectory of each leg.’
In the video, you can see the robot learning to navigate a changing environment whilst being manually agitated. The impressive work follows on from the lab's other great innovations with quadruped robots.