Researchers Develop a New Algorithm That Helps Robots Brace Themselves in a Fall

Now humanoid bots can preserve some of their dignity.

Oct 13 2015, 9:28pm

Say goodbye to the sad days of robots falling down, at least catastrophically.

Researchers at Georgia Tech have developed an algorithm that helps humanoid robots respond gracefully to several types of falls, bracing themselves differently according to the amount of force applied.

The video shows a humanoid bot responding to three different amounts of force, and compares it to the response of a bot that does not use the new algorithm. The team also measured the maximum acceleration of the robot's head, which was minimized with use of the algorithm.

In a press release, research team member Sehoon Ha explains: "Our work unified existing research about how to teach robots to fall by giving them a tool to automatically determine the total number of contacts (how many hands shoved it, for example), the order of contacts, and the position and timing of those contacts. All of that impacts the potential of a fall and changes the robot's response." Ha worked with Professor Karen Liu, who previously researched how cats brace themselves when falling. The two used that research to help develop the algorithm.

The research team points out that falls can cause serious and costly damage to a robot not properly equipped to protect itself. Hopefully humanoid bots will now be able to better follow Asimov's Third Law of Robotics.