In the realm of robotic canines, a new breed has emerged, boasting an artificial intelligence (AI) algorithm that empowers them with remarkable capabilities.
(Photo : Zipeng Fu YouTube Channel)
Parkour is a grand challenge for legged locomotion that requires robots to overcome various obstacles rapidly in complex environments.
Mastering Impressive Feats Through AI
Developed by AI researchers at Stanford University and Shanghai Qi Zhi Institute, Interesting Engineering reported that this cutting-edge vision-based algorithm endows these robodogs with the ability to conquer various challenges.
They can scale towering structures, leap across gaps, navigate beneath low-hanging obstacles, and traverse narrow crevices, all thanks to the robodog's algorithm, which acts as its cognitive center. Recent press releases from the respective institutions have shed light on this remarkable feat.
Chelsea Finn, an assistant professor of computer science, and the leading author of a peer-reviewed study on these robotic quadrupeds, conveyed her admiration for the project's achievements. She highlighted the remarkable autonomy and diverse range of complex skills that the quadruped robots have acquired.
What's particularly noteworthy is that these impressive results were attained using affordable and readily accessible off-the-shelf robot models, including two distinct ones. The newly developed robotic dog stands out from its counterparts due to its exceptional autonomy.
Search and Rescue Operations
This robotic canine achieves its autonomous operation through a combination of advanced perception and control systems. Equipped with a depth camera, it processes visual data using sophisticated machine learning algorithms.
As per Standford's official press release, this intricate interplay enables the robot to coordinate the movements of its legs effectively, granting it the ability to navigate diverse obstacles by going over, under, and around them.
These remarkable capabilities position the robotic dog as an ideal candidate for challenging search and rescue missions, particularly in environments with demanding conditions.
It excels at maneuvering through rugged terrains, accessing confined spaces, and aiding in the location and rescue of individuals in disaster-stricken areas.
Beyond search and rescue, its potential applications extend to military and defense contexts. Here, it can serve in roles such as reconnaissance, surveillance, and executing tasks in potentially hazardous scenarios.
Also Read: Boston Dynamics Robot Dogs Undergoes Training to Paint for National Gallery of Victoria
The remarkable aspect of our robotic dogs lies in their combination of vision and autonomy, which we can describe as their "athletic intelligence," as Fu explained. The development process was meticulously carried out, starting with the creation and refinement of the algorithm using a computer model.
This algorithm was then applied to two physical robodogs. These robots were given the freedom to operate as they saw fit and were subsequently rewarded based on their performance through a technique known as reinforcement learning.
This approach proved highly efficient in enabling the algorithm to progressively master the most effective strategies for addressing new challenges. The research team proceeded to conduct a series of extensive tests using real robodogs to showcase their innovative agility strategy.
Importantly, Hackster reported that these tests were conducted using the robodogs' existing computing resources, vision sensors, and power supplies. In terms of performance, the athletically intelligent robodogs demonstrated impressive abilities.
This includes scaling obstacles exceeding 1.5 times their height, leaping across gaps that were over 1.5 times their length, maneuvering under obstacles at 3/4 their height, and deftly navigating through narrow slits narrower than their width.
Related Article: [WATCH] Chinese Drone Flies with an Armed Robot Dog - Is it For Military?
