Making Robots Safer (To Themselves and Us)

DZone 's Guide to

Making Robots Safer (To Themselves and Us)

Making smart robots means putting an end to stupid mistakes...like robots that don't react to falling down. See what measures are being taken to make stumbling robots more graceful.

· IoT Zone ·
Free Resource

At a recent event, one of the speakers shared a story about working backstage at a Madonna concert. You probably heard about the gig, as it was one where the entertainer fell from the stage.

A long story short, he suggested that the fall was caused by a wardrobe malfunction, and Madonna actually knew a fall was inevitable and endeavored to take her tumble whilst doing as little damage as possible.

As a skill, it’s something that we thankfully don’t need all that often, but when it comes to robotics, it’s a skill that is largely lacking. And lets face it, robots fall over, a lot, so it’s not like it’s a skill that wouldn’t be rather valuable.

Falling Over Gracefully

A team of researchers at Georgia Tech are looking at a number of ways to help robots fall over slightly more gracefully.

The team developed an algorithm that allows the robot to work out how to manoeuvre its body as it tumbles so as to hit the ground with less force.

They put the algorithm through its paces at a recent conference in Germany using a humanoid robot called BioloidGP and in simulations of a large humanoid called Atlas.

The aim of the event was to try and simulate some of the challenges robots might face in real life environments, such as in failing nuclear power plants.

Suffice to say, trying to stay upright is arguably more important than ensuring any falls are damage free, but the high cost of each fall is incentive enough to try and mitigate the risks involved.

Some of the better approaches involved prompting each ‘limb’ of the robot to go limp when a fall was detected, thus hopefully dissipating any impacts incurred.

As technology advances making falls less likely to damage the robots themselves, the next step is to try and ensure they don’t harm any humans nearby.

The various methods to do this have thus far been limited by a lack of sensing capability in most robots. In humans, most of these tasks are undertaken by the nervous system, capable of reacting automatically to such challenges. With robots, however, everything has to be calculated, and with computational power still relatively low for such tasks, it represents a significant challenge to overcome.

One possible approach is being tried out in cars, where researchers are building algorithms that are able to predict our behaviors and thus are forearmed with suitable responses.

As robots become a more integral part of our lives, it will become essential that we can function safely in their company. It will be fascinating to see the various approaches researchers take to achieve this, and which ones eventually make it into the mainstream.

iot examples ,iot hardware ,iot world ,robotics ,robots

Published at DZone with permission of Adi Gaskell . See the original article here.

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}