Early this week, I read how Momentum Machines developed a fully functioning burger flipping machine.
And you know something, I was more surprised someone hadn’t built one earlier than I was about the new machine.
I mean, labour costs make up such a large portion of a company’s cost base.
Early last year, Japanese scientists developed a strawberry picking robot. The machine could pick about two thirds of a strawberry field in a single night. It did this by using three cameras to identify the ‘ripe’ berries.
Now I have no idea what the labour wages are for fruit pickers in Japan, but this seems like a very practicable labour-saving tool.
But not everything has to be practical. Not at first anyway. Sometimes the robots are just cool machines.
That’s the case with one of the eight robotic companies Google bought last year. Big Dog is a robot that can walk on ice, it can have its leg kicked out from underneath and keep going, oh, and it can climb 35 degree uneven terrain.
Or there’s Atlas. It’s a war like machine that has a fluid walking movement. Like Big Dog, Atlas can lose its balance but still stay upright.
If you haven’t seen the videos of these machines in motion, Google them. You’ll shake your head in amazement and wonder what they’ll think of next.
Oh wait. It’s here already. Meet iCUB. It’s named in part as an acronym of Cognitive Universal Body.
iCUB ‘playing’ with a ball
Ok, it doesn’t move like one of Google’s latest robotic toys. But the capabilities of iCUB are astounding.
This is possibly the most advanced humanoid robot in the world right now.
The iCUB, is a collaborative project across 25 different laboratories covering Europe, Japan and America.
Each of these labs have their own iCUB. The idea is that each lab does their own experiments with the robot and then shares the results. Other labs can then apply the new technology or continue to focus on their own work.
And that’s how the iCUB has become the testing ground for creating human conation.
Believe it or not, iCUB’s design is like a three year old child. It’s one metre tall, with childlike proportions. Scientists in the program have now found ways of getting iCUB to talk, walk, crawl, see and now it has a heightened sense of touch compared to other robots.
Overall, there are 53 motors controlling all of the movement. And about double that in joints to mimic life like flexibility.
The engineering alone is something to be amazed about.
But the real beauty of this machine is that it can learn.
Normally programmers spend an awful amount of time writing code to teach a robot how to recognise an object.
Instead, scientists have added extra senses to iCUB. Then programmers wrote code instructing the robot to use the senses to enquire about an object and then commit it to memory.
Basically, iCUB has been programmed to learn. iCUB acquires skills by exploring the surroundings using its body. It retains the data and then repeats the action.
Simply put, iCUB learns like a toddler – through experiment and repetition.
All of this learning is stored in the robots ‘memory’.
You can see the iCUB in action here. This is about as close as we get right now to artificial intelligence.
In the video, the instructor asks iCUB to do something. At first the robot doesn’t understand and tries to see how. The instructor then grabs the robot’s arm and completes the action with it. The robot then repeats the action itself.
iCUB has learnt how to do something.
This is remarkable technology.
I must say, when I first stumbled upon it I was amazed. I had no idea that robotics had entered this level of learning.
But what’s more surprising, is that none of this development in robotics would be possible without some very basic hardware.
Technology analyst Sam Volkering explained the role played by CPUs and GPUs in computing in the December issue of Revolutionary Tech Investor:
‘Your smartphone is a computer. If you strip away the screen and look beneath there’s a bunch of complex machinery whirring away. At the heat of it is the Central Processing Unit. This is important as it does a lot of the grunt work.
‘But as technology advances, and we demand more processing power from our devices, the humble CPU has its limitations.
‘That’s why the CPU’s best buddy is the Graphics Processing Unit (GPU). The GPU can process memory much faster than a CPU.’
Technology is advancing at a rapid rate, and so is the demand for technology that enables scientists and engineers to keep that rate going. Without the technology behind the CPUs and GPUs used in computing, the technological breakthroughs happening today wouldn’t be possible.
Whether it’s strawberry picking, burger flipping, artificial intelligence, or even virtual reality, it all needs ever increasing amounts of technology to make it happen.
And it’s the companies developing that technology that Sam is watching closely right now.
Shae Smith
Editor, Money Weekend
