AI robots: With the advancement in technology, the delivery services may be able to overcome the snow, heat, rain, and the gloom of night, but the new class of legged robots is not a way far from being achieved.
Artificial Intelligence and machine learning algorithms that are developed by the team of researchers from UC Berkeley, Carnegie Mellon University, and Facebook are trying to equipped AI robots with legs and that too with an enhanced ability to adapt and navigate an unfamiliar piece of land in the real meantime.
During the test, their robots successfully traversed sand, tall grass, mud, hiking trails, and dirt piles without falling for a single time. The system has also performed extremely well as an alternative system in adapting to a weighted to throw a backpack on its top or too slippery and oily slopes.
The new robots with the help of AI can map the ground while walking down steps and moving over the piles of cemented floor or pebbles has achieved a success rate of approx 80 percent respectively but still, there is an impressive feat that has given the lack of simulation calibration for prior experience with the instability in the environment.
The new AI and Machine learning has enabled the new era of legged robots which consists of Rapid Motor Adaptation that was enabled in test robots traverses of various types of terrain. The robot not only could adjust to the novel circumstances, but it can also do similar in fractions of a second rather than in a minute or more. This can be critical during practical deployment in the real world.
In a study principal investigator of Rapid Motor Adaptation, Jitendra Malik that their insight is that to change is ubiquitous starting from day one and the RMA policy assumes that the environment will be new.
AI robots, He is a professor at UC Berkeley’s Department of Electrical Engineering and Computer Sciences and a research scientist at the Facebook AI Research group. The RMA project is a part of an industry-academic collaboration with the FAIR group.