Edge of Innovation

Robots that look like humans? These designs are far more efficient

 

The new Channel 4/AMC show Humans paints a picture of robotics where human-like synths ultimately create more problems than they solve. It’s science fiction, playing upon our fears of technology going off the rails.

While we’re a million miles away from being able to create a do-everything android helper that looks like Gemma Chan, robots are coming. Slowly and awkwardly, perhaps. And they’re none too pretty. But they’re coming. Forget dumb manufacturing machines, kid’s toys (that don’t do as much as the box says they do), and promising-but-impractical lab experiments. Robotics in 2015 is a whole different and quite surprising beast.

Take, for instance, the snakelike robot that was used to assess the damage inside Japan’s Fukushima Daiichi nuclear plant after 2011’s disastrous radiation leak rendered the surrounding area unfit for humans.

Two feet long, the clever Hitachi robot was able to enter the damaged plant by way of a thin pipe, rearranging its segmented body into a “U” shape to move through the shattered building — all the while recording images and radiation readings for the robot’s human operators.

Amazon already uses Kiva robots in some of its warehouses to bring shelves to its workers. Image copyright Amazon.com.
Amazon already uses Kiva robots in some of its warehouses to bring shelves to its workers. Image copyright Amazon.com.

In the US, shopping giant Amazon has been trying to automate its huge warehouses. Rather than have its human workforce pace the aisles to find the shelves (and the products) to fulfil orders, the company uses a fleet of bright orange Kiva robots to move the shelves to its workers.

Even in this scenario, while eliminating walking improves efficiency, Amazon is still forced to rely on human employees to locate and ‘pick’ the products from the shelves. But maybe that will change in the future.

Amazon ran a competition earlier this year to find a robot capable of not just moving products around its massive warehouses, but intelligently picking up and boxing items from its shelves. The competition was won by a robot made by the Technical University of Berlin, which was able to use a suction cup and special sensors to recognise objects and carry out much of the functionality that Amazon hopes will eventually replace its fleshy workforce.

A similar new project carried at U.C. Berkeley utilised machine intelligence to create a robot able to learn tasks through trial and error, much like humans do. Instead of building specific robots for specific tasks, U.C. Berkeley’s BRETT (Berkeley Robot for the Elimination of Tedious Tasks) demonstrates that artificial intelligence and deep learning algorithms can allow standard robots to pick up new skills within minutes.

If speed is your thing, of course, you may want to turn your attention to MIT’s ultra-impressive Cheetah robot. Sprinting and jumping over hurdles like the big cat that it’s based on, this quadruped machine can autonomously leap obstacles more than 15-inches in height as it bounds across the floor. You wouldn’t want one of these chasing you down.

You might think that the cutting edge of current robotics technology would be one of the robots that competed in the 2015 DARPA Robotics Challenge. Twenty-three automatons tackled disaster-themed courses featuring eight complex tasks that included driving a vehicle, climbing a set of stairs and cutting a hole in a wall.

The competition was won by South Korea’s Team KAIST and its DRC-HUBO picked up the $2 million grand prize, completing the tasks in under 45 minutes. The Running Man robot devised by Florida’s Institute for Human and Machine Cognition came second, while team Tartan Rescue from Pittsburgh finished third with its robot CHIMP.

Team VIGIR DARPA Robotics Challenge Finals 2015
Team VIGIR’s robot negotiates a door in the DARPA Robotics Challenge Finals 2015.

Of course, while Humans fashions a future where robots look like us, most robots don’t. Or won’t. Tomorrow’s robots will be delivery drones and driverless cars, home automation systems and nano-sized body monitoring devices.

The last few years have seen a massive breakthrough in robotic agility which is making possible new things previously unimaginable to researchers. This will have enormous positive implications for humans. For example, Robert Gregg of the University of Texas’ Erik Jonsson School of Engineering and Computer Science has led a project to create wearable robots, which one day might help disabled people to walk again.

“The recent technological breakthroughs in autonomous walking and running robots are beginning to be used in wearable robots that could restore mobility in persons with physical or neurological impairments,” Dr. Robert tells IQ. “Wearable robots are still catching up with human performance in 2015, but they might one day exceed it.”

Whatever their applications, it seems that robots are here to stay. We don’t need them to look like us to be helpful. — Luke Dormehl (@lukedormehl)

Share This Article

Related Topics

Tech Innovation

Read This Next