LESSON

Robots today

Transcript

If you ask people today to say what robot looks like, they're either going to tell you about a fictional robot, like R2D2 or they're going to mention a manufacturing robot such as the one shown here and there are many many of these robots at work on the planet today, more than 1 million of them and they underpin a lot of modern manufacturing, of automobiles, of electronic devices, like computers.

They also do a lot of material handling, moving boxes off conveyor belts and putting them in shipping cartons, even picking up shipping cartons and putting them onto pallets and so on. Robots such as this represent the first generation of robots. They are really direct descendants of the robots created by Unimation Inc. that we talked about in the last section.

Here's an example of a modern manufacturing robot at work. This is a video that I took at a robot exhibition last year. And what we can see is the robot is moving and picking up a piece from the bucket at the bottom of the ramp, and you can see that the pieces are all in different orientations and the robot has got some sensors so that it can determine the orientation of its gripper that it needs to adopt in order to pick-up the object, and once it has done that, moves it to the top of the ramp and let's it go. This robot is performing and endless loop of pick up the thing from location A and drop it at location B.

Another type of robot is this. This is the robot vacuum cleaner and this is a technology that was introduced around 2004 and since that time they became enormously popular, more than 10 million of them have been sold. So the manufacturing robot that we saw a moment ago is a technology that's nearly 60 years old and there's a million of them at work. This technology which is only just over a decade old has sold more than 10 million of them.

One of the reasons these robots are so popular is they perform a very useful task, a task that human beings don't like doing so much, and they do it for a very low price. The reason that they're able to build a robot for such a low cost is it has got really limited computational capability and robots are able to clean the floor by adopting a very very simple strategy. The robot doesn't actually know where it is. What it does is just drives in a straight line, sucking up dirt until it hits an obstacle and bounces off the obstacle and then goes in a different direction.

So this time lapse picture here, shows the path of a Roomba robot cleaning a living room and we can see where it's bounced off the walls and off the furniture. By adopting this very simple strategy, we can cover the whole room. It's clear though, that some parts of the room are covered many many times and some parts of the room are covered not very often at all.

We can imagine scaling up a technology like that vacuum cleaner to a robot system that's able to pick up and put down containers in a shipping port, and in fact people have built remote systems to do this.

So these robots pick up a container at location A and transport it to location B, when they put it down again. Robots that move materials from location A to location B, inside manufacturing plants are now really quite common, and this separates them from the first generation of robots. The manufacturing robots which are fixed at a single place in the world, the next generation of robots are able to move within the environment.

Here we see a research robot called little dog, and it played a really important role in research into how we plan where robots place their feet in order to walk competently over very rough and uneven terrain.

Here we see the big brother of little dog, logically called big dog, and this is a robot that's able to carry quite a large payload over very rough terrain in realistic out door environments. Here we see a demonstration of the stability of this robot walking over ice and being kicked in the side and able to recover its balance and keep on walking. Once again we can see wonderful demonstration of its ability to recover from very very difficult situations.

In addition to wheels and legs, robots can also fly and people have demonstrated very large scale flying robots like the Global Hawk here on the left and really, really small flying robots are shown here on the right.

So what's the definition of a robot? A definition that I like a lot and I think is actually very very useful, even though it's perhaps a little simple is it's a machine that moves things from A to B. Now, it's a programmable machine so we can control the location A and the location B and perhaps the location A is a function of information that the robot picks up from its sensors. But in the simplest case, a robot is a machine that moves things from A to B.

From mobile robots such as those that got wheels or legs or propellers and can fly, we can think of it as a machine that moves automatically from location A to location B. A high level, a more complex definition of a robot that I like is it is a goal-oriented machine that can sense, plan and act, and this phrase catches, I think the four really key elements of a robot.

The first thing is, it's a machine that has a goal. There's something that it wants to achieve, it wants to deliver itself from one place to another or it wants to move something from one place to another, so that is its goal.

In the example of the manufacturing robot, we saw that it was able to sense where are the white objects that it wanted to pick-up. We saw with a little dog walking robot, it was able to sense the shape of its environment in order to determine where its feet should go and that brings us then to the next part, it's to plan. So once the robot has sensed where it is and sense something about the environment in which it's moving, then it makes a plan to bring the robot closer to its goal and once it's created a plan in order to move closer to its goal, then it acts.

It moves the wheels, it moves the legs, it changes the speed on the propellers, in order for the robot to get closer towards the goal. And it repeats these three actions continuously. It senses, it plans, it acts. It senses the state of the world again and it updates it plan and then it acts a bit more and so on until it reaches the goal. What are the characteristics of robots? Why are robots useful to us? Well, they have a number of merits. First of all robots are very consistent, they pay attention all the time, unlike human beings.

Robots are also very accurate. The manufacturing robots that we saw earlier are able to position the end of the robot to an accuracy of a fraction of a millimeter, every single time they perform the operation and they're also very, very reliable. They can work 24 by 7. They don't even need the lights on in order to do their job. So this means that robots are able to do a number of things that people can't do. They can go into space, they can go deep into the oceans where human beings are just not able to survive. They also do things that human beings won't do. These are jobs that are boring or dull.

And increasingly there are a lot of low-skill jobs in agriculture and in manufacturing where it's very difficult to find human beings willing to do that work. These are jobs human beings now consider are beneath them and won't do. Now robots are going to move into that job vacuum. And finally, robots do things that people shouldn't do. They can perform jobs that are dangerous, unhealthy or risky, that is, the robots take people out of harm's way and I think that's a very great thing that robots can do.

People often use the expression, 'dirty, dull and dangerous' to reflect the three classes of jobs where robots have considerable advantage over human beings. The dirty and dangerous jobs are the ones that human beings shouldn't do and the dull jobs are the ones that human beings don't want to do.

Robots today are ubiquitous in manufacturing but they can do much, much more.

Professor Peter Corke

Professor of Robotic Vision at QUT and Director of the Australian Centre for Robotic Vision (ACRV). Peter is also a Fellow of the IEEE, a senior Fellow of the Higher Education Academy, and on the editorial board of several robotics research journals.

Skill level

This content assumes only general knowledge.

More information...

Rate this lesson

Average

Check your understanding

Leave a comment