Exploring the Agility of the ‘Parkour Robot Dog: Leaping, Jumping, and Crawling Capabilities Unveiled

Parkour Robot Dog

Parkour Robot Dog we have already seen that four-legged, dog-inspired quadruped robots are capable of a wide variety of both entertaining and useful tasks including remotely monitoring sports stadiums and guiding the blind to inspecting potentially dangerous research areas. Typically, these robots are more agile than their hulking bipedal counterparts, but few of them have come close to matching the fluid grace and undeniable athleticism of their furry canine inspirations. Now, though, a new wall-climbing robot is stretching our conception of what these quadrupeds are capable of, and it is doing so with considerable style.
The researchers from ETH Zurich are working to close the gap between robots and animals. They developed a mobile robot capable of running, jumping, and crawling through obstacle courses.
“This work is about closing the mobility gap between robots and animals,” researchers, who published their report in Science Robotics , said. They taught 100 pound quadruped robot ANY mal, made by the ANY botics to mimic people who practice an underground sport. Many call it Parkour Robot Dog, which they termed ‘freerun.’
To put it briefly, parkour is all about reaching the destination as quickly as possible .
Therefore, one has to crawl quickly and jump through any obstacles available on the way . Such a strategy requires combined athletism while making fast decisions. ANY mal coped with the task easily. The robot, which was greatly improved in terms of quality and other characteristics, was able to complete the specified Parkour Robot Dog course moving at 6 ft/s pace.
A video of brand-new ANY mal in action from ETH Zurich is as satisfying as it is oddly fantastic. This beefy red robot methodically climbs up a small wooden staircase and then leads over a small gap to fall and land on another small, high table. Without slowing down, the robot takes a bold step forward and throws itself forward and then dives under the obstacle licking reminiscent of some of the more searching insects. And then it pushes itself back to get up and go vertically along another high box. The filmed robot successfully completed the route, even when the obstacles were arranged differently.
The robot ANY mal is equipped with onboard laser sensors with the help of which it perceives the world around it, builds maps on the basis of them, and independently plans and follows the path to the destination. A total of four lightweight carbon legs and 12 identical motors move the robot in the right direction. To improve their functionality, ETH scientists used a neural network, which consists of three modules, one of which is responsible for movement, the second – for vision, and the third – for the path search.
According to the paper, the researchers created the navigation element of ANY mal to determine the robot’s ability to walk, jump, and crouch. With that information in hand, ANY mal is now able to adapt its behavior autonomously depending on what kind of obstacles it is prevented. As a result, the robot is able to instantaneously detect and respond to a variety of obstacles and move over them . Nikita Rudin, an ETH doctoral student who participated in the work on the improvement of ANY mal, is a Parkour Robot Dog himself, so he drew lessons from his past experience.
“Some of my researcher colleagues thought that legged robots had already explored all they could before we started this project, but I had a different opinion,” summed up Rudin, as cited in Science Daily . “I, in fact, was absolutely sure that a lot can be done with the mechanics of legged robots.”
The model was trained on many examples of human free runners who can be seen crossing Parkour Robot Dogobstacles. During training, ANY mal learned step by step in simulations of various difficulty and design. So the robot could go through real obstacles, which was prepared with small physical differences .
In a statement, the research says, “By aiming to match the agility of free runners, we can better understand the limitations of each component in the pipeline from perception to actuation, circumvent those limits, and generally increase the capabilities of our robots, which in return paves the road for many new applications, such as search and rescue in collapsed buildings or complex natural terrains.

• Parkour Robot Dog skills might help robots designed for search and rescue and space exploration

ANY mal’s new tricks were developed with practicalities rather than aesthetics in mind. The researchers believe that in the next decade, similar abilities will be added to search and rescue robots to help them jump over fences or run up stairs to reach difficult-to-access places in case of emergency. Later on, according to the researchers, these abilities might help the robots conducting the space mission walk on extremely rocky surface of the Moon or another planet. The new functionality can also be used by quadruped robots crawling through the Large Hadron Collider in Geneva, Switzerland.
Many of these use cases are still largely speculative. In the meantime, however, ANY Mal has joined the Boston Dynamics’ Spot and Atlas robots in landing on the ever-growing list of metal contraptions that can perform athletic stunts that before seemed to be reserved for sentient beings.