About a decade ago, the Pentagon's Defense Advanced Research Projects Agency organized a contest for robots to try out tasks like opening doors, using power tools and driving golf carts.
Jerry Pratt is a senior scientist at the Institute for Human and Machine Cognition in Florida. He led a team that finished second at the 2015 DARPA Challenge. Currently, he serves as the CTO of a graphics artificial intelligence company that specializes in designing humanoid robots for warehouse tasks. Recently, the company announced a $70 million investment**. Pratt noted that if the DARPA Challenge were held today, the robot would probably take only a quarter of the time it would take his robot to complete the task at the time, and the probability of an accident would be much lower. He stressed that a number of new technologies have emerged that have greatly improved the capacity needed to meet these challenges.
Advances in computer vision are largely due to advances in machine learning over the past decade, which have significantly improved the machine's ability to maneuver in complex environments and perform actions such as climbing stairs and grasping objects. In addition, the development of high-energy-density batteries and the advancement of electric vehicle products have provided sufficient power for humanoid robots. This power allows the robot to move its legs quickly and maintain dynamic balance in the event of a possible slip or miscalculation, just like a human.
Pratt mentioned that his company's robots are starting initial operations in a simulated warehouse in Sunnyvale, California. Brett Adcock, CEO of Graphics, believes that it is feasible to produce humanoid robots at a cost similar to that of building a car, given the market's demand for increased production. This view implies that robotics is approaching an important milestone. You're probably familiar with the dancing Atlas humanoid robots that have garnered a lot of attention on YouTube. The robots were manufactured by Boston Dynamics, a company known for its pioneering work on leg movements and for creating some humanoid robots in the Defense Advanced Research Projects Agency (DARPA) competition. They demonstrated the possibility of building capable humanoid robots. However, these robots** are expensive, initially costing millions of dollars, and lack the necessary software to operate automatically and really function.
In addition to graphics companies, there are other companies that are also working on humanoid robots, including 1X, Apptronik, and Tesla. Tesla's CEO, Elon Musk, participated in the Defense Advanced Research Projects Agency (DARPA) Robotics Challenge in 2015. He is currently showing a keen interest in developing humanoid robots, which means that some of the key technologies needed to implement such a machine are becoming viable. Jonathan Hirst, a professor at Oregon State University and co-founder of Agile Robotics, also took part in the DARPA challenge and showed off the walking robots he designed. Agile Robotics has been working on developing robots with legs, but Hirst mentioned that they focus more on physical movement than on mimicking the mechanics of human limbs. So, although these robots look like humans, their legs are more like ostriches.
In March, at a manufacturing conference called Promat, Agile demonstrated the role of its robots in a warehouse. These responsibilities include independently picking suitcases from shelves and then placing them accurately on the conveyor belt. Currently, many warehouse and manufacturing robots use wheels instead of legs. In addition, traditional automation methods such as conveyor belts can also be cleverly integrated into the structure of the warehouse.
However, Meloneewise, CTO of Agile Corporation, highlighted the superiority of leg systems in many situations, especially in companies that can't fully automate operations. Wise says humanoid robots have distinct advantages: they are more adept at navigating stairs, ramps and uneven surfaces;More flexibility in tight spaces;Perform tasks that bend or stretch, which are not available in wheeled robots. Wise has recently advocated for humanoid robots and was once the CEO of Fetch Robotics, a company that specializes in wheeled warehouse robots. She said the market was ready and highlighted the imminent challenge of improving reliability, emphasizing that in robotics, success depends on the ability to handle failure gracefully. Despite their clumsiness, the robots from the DARPA challenge were significantly ahead of their time.
If you are interested in the article, please follow me or leave me a message