Cracking the Code

Being a racing driver is an exciting job

that needs split-second reactions and nerves of steel.

When driving round tracks at high speeds,

it's important that drivers know exactly what they're doing.

Nowadays, racing simulators are used by all the top racing teams

to train their drivers.

These simulators aim to recreate the racing experience

as closely as possible.

So, how do computers make the virtual world so realistic?

I'm here at Silverstone racing track

with my computer hackers to find out more.

I'm Bradley.

I'm Hannah.

Hi, I'm Stuart, I'm going to be your driver.

Before we can find out how realistic simulators are,

we first need to know what it's like in a real car.

So we're getting in the back

as racing driver Stuart takes the wheel.

Wow, that is so fast and really scary!

Luckily, Stuart's a professional who knows what he's doing.

To get that good, he's needed to do a lot of practice.

OK, how was it?

-Amazing.

-It was really cool.

How come you can go so fast on the track?

It's fair to say I've probably done quite a few laps around Silverstone,

I do know it very well.

And also I know the car very well,

so you put those two things together

and obviously that makes it easy for me.

But nowadays, drivers don't do all their practice on the real track.

Our next stop is the University of Hertfordshire

to see how their racing simulator helps drivers

to get the perfect lap time.

I'm here with Geoff, who runs the simulator,

and Claire, whose company makes the simulators.

So guys, what do you think?

Awesome.

Absolutely amazing.

What's it made up from?

The simulator is a complex mix of high-tech hardware and software.

On the hardware side, we've got the motion base,

we have the steering wheel, the pedals, the screens.

And on the software side,

we have lots of models working hard behind the scenes

to make sure the driver sees the right thing

and feels the right thing.

And to make sure the driver does feel and see the right thing,

programmers have to know about physics.

They need to understand how cars behave in real life,

so that they can programme the simulator to work in the same way.

-So if you practise on one of these...

-Yeah.

..would you be able to drive a real car?

You would, yeah.

I mean, drivers come in here, professional drivers,

they will drive the car, learn the circuit

and we've had drivers win the race straight after this,

they've never been to the circuit before.

Can we have a go?

Of course, let's go and do it.

In order to make the simulator as realistic as possible,

programmers have to make a very accurate virtual world

inside the computer.

We call this modelling.

So the simulator code must use the laws of physics,

it has to know how to work out the speed of the car

or the forces that the driver feels.

It also needs lots of information - lengths, heights, weights,

numbers that describe the car, the track, the bumps, everything.

So we create our model of the world using rules and information.

This file describes the car,

so if we have a look at it here,

there's various numbers,

you've got the mass of the car,

so that tells you how much the car weighs in kilograms.

We have power, we have braking,

we have the torque of the engine,

so how much power the engine's able to deliver to the track.

What happens if you change the numbers?

These numbers?

Well, you could make a new car.

If we took some of these numbers,

maybe we changed 20 of the numbers,

things like the engine performance, the engine speed,

the weight of the car...

we can make that Renault Megane drive like the F1 car

and be as fast as the F1 car.

That's pretty cool.

So can we have a go with the Formula 1?

We can, yeah, but that is tough,

you've got to be good to drive the F1 car.

So Hannah and Bradley had a few crashes in the Formula 1 car.

Good job it's just a simulator.

But now, it's time to find out how a professional does it.

We've seen how we can represent different cars using information

like the mass of the car or the power of the engine.

But how does a simulator use that information

to create a thrilling experience which responds to the driver?

That's done by writing the computer program.

How do you programme the simulator?

What we can see here is the core computer code of the simulator,

we can see it's a set of many, many commands -

if you turn left, the simulator will move in such a way

that you feel like you're really turning left.

If you put your foot down on the accelerator,

it's going to make you feel you're going really fast.

That's all written in this code.

On screen, we've got a load of words and symbols,

but when we run this code,

it creates the whole simulator experience

and brings the virtual world to life.

This code is where the programmers are applying the laws of physics.

What else does the simulator have to do?

The simulator is having to do so many different things at once.

If you're the driver and you're hurtling towards a corner,

the simulator needs to know -

are you going to make the corner in time?

If you don't, are you going to spin off and hit the wall?

Is it raining? Are you going to slip?

What noises does it need to make, what do the graphics need to show?

Every single element of the simulator is talking to each other

-about 1,000 times a second.

-Wow!

So, guys, what did you think of that?

-Awesome.

-Scary.

So, how did the two of you think it compared to Silverstone?

When you were racing round, it looked really real.

It was fun,

but it wasn't quite as real as being actually at Silverstone.

The simulator isn't perfect,

because there are limits to what computers can do.

So how much information they can read,

how fast they can work out calculations,

or how much detail the graphics card can draw on the screen.

So simulators may still have some way to go,

but computer programmers are improving on their programs,

and newer computers are making it easier and faster

to work with more information,

so simulators can become even more realistic,

to the point where maybe it will be difficult to tell the difference

between computer-generated environments and the real world.

Video games are popular all over the world

and I know I can be found playing on my games console for hours.

But creating the game can be just as fun and challenging

as playing the game itself.

So I've come to St Saviour's Primary School, in London,

to meet the year sixes who are using Scratch to create games.

So do you all like playing video games?

ALL: Yes.

All of you, that's great.

Well, today, you are going to get some cool stuff,

you're going to create your own video game.

It's a simple game of cat and mouse,

and you're controlling the mouse with your mouse.

'The children are using a free piece of coding software.

'All the commands you need are laid out on the left hand side,

'these are all the instructions which make up your code.

'You then drag all the commands into the middle to write your program.

'On the right, you have your characters, or sprites,

'and above this, your stage, where you can view everything.'

-The mouse is trying to run away from the cat.

-Oh, wow.

OK, so how are you doing that?

-You have to give it a script.

-OK.

So for the mouse mine is "move 20 steps"

and it would be quite boring

if you had to press it and press it over again,

so you would go to Control,

then you would put "for ever",

-or you can do "repeat", but I'm choosing "for ever".

-OK.

And so you could put sounds on it,

just jam it in through here

and it has to, you have to go to Controls again,

and there's an option saying "when green flag clicked",

so you put it at the very top.

And let's test it out and...

Oh yeah, that's cool.

'Although the children are designing the same basic game,

'the beauty of using coding software such as this

'is that it gives the freedom to develop your own version

'and be as creative as you like.'

This mouse, Herbert, follows me around and the cat, Felix,

follows Herbert around,

but when Herbert gets caught by Felix, he turns into a ghost.

Wow!

'Like all good games, we need a scoring system.

'A score is a variable, so you need to use the variable commands

'in the top left of the screen.'

Every time Herbert gets caught by Felix,

the score goes minus 100,

and it goes up every second

when you keep him away from Felix.

OK, so every time the mouse gets caught by Felix,

-you lose 100 points.

-Yeah.

But, as long as you keep the mouse away from Felix,

-you're gaining points all the time.

-Yeah.

-So you have four sprites in your game?

-Yeah.

That's a bit complicated.

-Yeah, it's a harder game to play than the one on here.

-Uh-huh.

I still need to work a bit on the programming of the two new ones,

cos it's not really working with the other two ones.

Yeah, at the moment, it looks like Herbert and that fish,

-one of your new sprites, is kind of glued together.

-Yeah.

All right, well, it seems awfully complicated,

so I'll let you get on with that.

'When you're programming, things often don't work out the first time,

'so you have to really make sure you've understood the problem

'and that you have all the right steps, and in the right order.

'All this coding has got them thinking

'about the kind of games they'd like to create,

'the rules of the games, and how they'd fit together.'

I would like to make a game based on pinball,

and when the ball touches some holes, it gets points.

I would like to make a maze

and you have to... The character will follow you with your mouse

and you have to collect like, food items along the way to power up.

I would like to make my own car game.

You have to avoid objects

and if you hit an object, you lose points.

Now you've heard their ideas,

what sorts of game would you make?

Where are we?

The Parliament of the Daleks.

I don't know about you,

but I'm a massive Doctor Who fan.

I grew up watching The Doctor,

transfixed as he travelled around the universe

escaping from all the weird and wonderful cultures.

A lot's changed since then,

different Doctors have come and gone,

but I think it's fair to say that the show gets better and better

and looks more and more stunning.

And a lot of that is due to computer programming.

Hi, I'm Joey.

Hi, I'm Effie.

So me and my two lucky hackers, Effie and Joey,

who are both Doctor Who crazy,

have both come to BBC Television Centre

to find out where it all began.

Back in the 1960s, when Doctor Who started,

they used to make all these props by hand,

I mean, they still make some of the props by hand now,

but it's all a bit time-consuming,

so how else do you think they make them?

-Using computers...

-Graphics and stuff.

Shall we go and find out?

-Yeah.

-Come on, let's go, then.

The next stop is the Visual Effects Studio.

There are many elements that go into making Doctor Who

such an amazing, magical show.

But some of it isn't possible with just actors and props.

So, guys, we're here with Cat and guess what?

She works on Doctor Who!

Really?

Do you actually do the episodes and stuff and like that?

Yeah, we work on a lot of the Doctor Who episodes.

Cat shows the children

how computers are used

to produce really exciting visual effects for the series.

So, let's have a look at this Dalek,

this is one Dalek by itself, but...

-Lonely.

-Yeah, he looks lonely.

So how do you think we could make a bunch more Daleks?

BOTH: Um... Copy and paste?

THEY CHUCKLE

Well, we could do that, but it would take a really, really long time,

so, instead, we're going to be lazy and use some programming.

And I've got this little bit of script here

and I can run that to make a set of Daleks,

six rows with six in each row.

'So with this piece of code,

'we can create six rows of six Daleks from the one original Dalek.

'The code tells us that for each row,

'and for every position within the row, make one Dalek.

'Then, move the Dalek into position.'

So, if I just run that.

Wow! That's a lot of Daleks.

We have a little army of Daleks.

When I watched Doctor Who as a kid,

often from behind the sofa,

you could only ever have a few Daleks on screen,

because they were actual props,

but here we're creating whole armies with a little computer code.

'It's terrifying.'

What's even more amazing

is that every single Dalek in that army can behave independently.

Using a piece of code, called the rand command,

we can instruct each Dalek to move its eyestalk,

head and body randomly.

So can we all have a go at changing the code?

Yeah, sure, we could make them look in the same direction,

if you'd like.

Yeah, that would look cool.

So let's change the head rotate,

cos that's which direction they're looking in,

so, instead of this, which picks a random number,

what we're going to do is we can put in just a number that you choose.

So let's give them a number.

Um... I don't know.

Pick a number.

By changing the code, we replace the rand command with a set number

and hey presto, you have an army of Daleks staring in one direction.

And they all look the same,

exactly the same way.

So you wouldn't want to be facing them.

'But that's exactly what we're going to do

'as visual effects artist, Rose,

'puts our acting skills to the test.

'Right, Rose, what's our motivation?'

In this scene, it's the Dalek Parliament,

so you're all surrounded by Daleks,

and I want you to look like, kind of scared.

Brilliant. Yeah, that works, like that.

Brilliant. And action.

When we want to mix actors with computer-generated effects,

you need to film them against a green screen,

so you can replace the background

with something you've designed yourself.

They're everywhere, what are you going to do?

Right, now, stand up, stand up.

Cool, now, look really brave.

Now, you're not going to be scared of these guys,

the Doctor's going to come along and save you in a bit.

'OK, that's a wrap.

'Now, it's back to the edit to upload and review the footage.'

You look great, don't you?

Your face is all covered up.

Wow, look at you, guys.

Oh, it's hilarious!

So shall we put you guys in this?

All right, so...

I'm going to use the computer here

to select all of the green and get rid of it.

-Are you ready?

-Yeah.

It's going to look really cool.

There it goes, see?

Is this how they film Doctor Who, then?

Yeah, this is how all of the green screen shots are done on Doctor Who,

using this kind of method.

That's so cool!

How long does it take to film an episode?

It depends on the episode,

it can be anywhere between a few weeks

or a couple of months, it depends on how many creatures are in it.

Now, I can put you on top of the footage I have here,

so that you'll be sitting in the Dalek Parliament.

So here we go.

That's you guys there.

Wow. That's so cool.

'Rose has used a graphics package to superimpose us into the shot.

'The computer runs a program

'which has removed all the green pixels from the footage

'and then merged the video with the computer-generated images.

'Hollywood, eat your heart out.

'Here comes the finished shot,

'the three of us transported into the Dalek Parliament.'

Here I am!

'And here's the Doctor on the same set.'

Remember, next time, when you watch all the monsters on Doctor Who,

that lots of them were generated by computers

and computer code was used to make them move around and look scary.

And it's all done by artists and coders

sitting around computers like these.

When writing your own computer program,

you'll sometimes want to include graphics.

I'm here with the pupils from the Ilsleys Primary School

so we can create our own computer graphics.

You can create something known as a bitmap

and that's when you break the picture down into a grid

of different coloured squares.

Hi, guys!

So we've given you a stack of cards and also a grid, known as a bitmap,

and that has a series of zeros and ones.

Every time you see a zero, that represents white.

When you see a one, that represents a black.

So just turn over the cards when you see a one

and see what picture you get.

'We've got two different teams,

'and each team is working through a different bitmap grid.'

Zero up there, white up there.

'When there's a zero in the grid, they place the card white face up,

'when there's a one, they place it black face up.

'We're using only zeros and ones to represent the information

'and we call this binary.

'So a computer can represent graphics and pictures

'as a grid of numbers.

'Each number tells us the colour of one small square called a pixel,

'and we build up the whole picture using a load of these pixels.'

All right, so, guys, how are you approaching this,

how are you working as a team?

We're doing the corners first and then seeing where they're going.

-So you're the director, telling everyone what to do.

-Yeah.

And they place them. Oh, that's cool.

Guys, wow, you seem to be doing really well. What's your strategy?

-Going row, row, row.

-Row, row, row.

Cool, so what do you think it could be?

I have no idea, it looks like antennae.

-We don't know...

-You've lost where you're at.

One, two, three, four, five, you're on the sixth level up.

Wow, you guys were super quick, that was awesome.

So what do you think it could be?

-Alien.

-An alien? Do you all think that?

-Like a robot.

-A robot?

What about computers?

Think about games.

-Space Invaders!

-Space Invaders, I think that's right.

Wow, you guys have done it as well.

So what do you reckon this is?

-A dog.

-A dog? What makes you say that?

Because it's got the legs and it's got the ears,

and it's kind of got the body shape.

-OK, oh, yeah, and that over there's a tail, I suppose.

-Yeah.

OK, no eyes, though.

But I've got to ask you, how come that leg down there,

it seems by itself?

Is there a black missing?

So I reckon this one actually should be a black,

I think it was a number one.

Yeah...

-Now, that looks a bit better, doesn't it?

-Yeah.

So, guys, these look totally awesome.

However, they do look rather blocky

and that's because the squares we're using are rather big.

So I have some smaller squares, which are a quarter of the size -

you can fit four of these

into one of the squares you've got at the moment

making it a bit smoother, and that's something we call higher resolution.

'A higher resolution picture contains more pixels,

'so there's more information and we can have more detail.

'And there are other ways to improve the picture too.

'Today, we're only using zero and one for white and black,

'but if we used more numbers, we could have different colours.

'When your computer stores a photo,

'each pixel could be any one of over 16,000,000 colours.'

That's better already, can you see? It's actually less harsh,

it curves a bit more, these edges are a lot smoother.

Yeah, that's working really well, cool.

This is pretty impressive.

Oh, you've even put the white bits in the ears.

Yeah, I thought if we could have even smaller squares,

coloured maybe pink inside,

it would look even more effective.

Yeah, you could have the pink inside the ears, that would be great.

So what do you think, do you like it, do you think it looks better?

A bit smoother?

It has a higher resolution.

So, guys, these look great.

The smaller squares have improved it already,

and if we were to go even smaller with the squares,

we could make it much better, and if we were to add more numbers,

not just the zeros and ones, we could actually add some colour too.

Nowadays, miniature computers are everywhere,

inside our tablets and mobile phones.

This is a new generation computer,

the size of a credit card.

These tiny computers are what my hackers and I

have come to see today.

I'm Lottie.

I'm Harry.

And this is Dave.

Today, he's showing us the exciting things

we can do with a tiny computer and a bit of programming.

What is a Raspberry Pi?

It's a small computer, it's very cheap and it's very lightweight.

I thought it would be about that big,

but it's a lot smaller and it can still fit everything you need.

Some things you recognise off a normal computer,

and there's other things you've never seen before

because it's like on the inside, it's inside a computer.

Because it's a fully functional computer,

you can programme it to do all sorts of amazing things,

like building your own electronics projects.

And because it's small,

Dave and his friends have done something really cool.

You might think these photos were taken by satellite,

but, actually, they were taken from this balloon.

Hanging from the balloon is a box which contains an ordinary webcam,

a sat nav tracker, a radio transmitter

and, of course, the miniature computer,

which Dave has programmed to snap the photos

and then beam them back down to Earth.

So this is what we have in the payload, as it's called,

there's the webcam that it takes photos with,

and that's the aerial, so that dangles out the bottom

of the box and that transmits signals maybe 300, 500 miles away.

And how do you get the signals?

I've got a ten-metre pole in the garden there

-with an aerial up the top.

-Oh, wow!

'So Dave's connected the webcam and sat nav tracker as inputs

'and the radio transmitter is the output.

'Now we know how it all works,

'let's get the kit ready for launch.'

Perfect, and then just slide in.

'The computer and webcam go into the protective box,

'and this will be hung underneath the balloon.'

That's probably the biggest balloon you've ever seen.

'Dave fills the balloon with helium, just like a massive party balloon.'

It's attacking me.

SHE LAUGHS

'With Dave and his friends, we ease it up into the sky.

'There's a balloon, a parachute,

'our computer and a backup transmitter

'so we can find it even if the computer goes wrong.'

Three, two, one!

THEY LAUGH

Bye-bye, balloon, see you later.

With the balloon launched,

the webcam is already taking pictures.

Each picture is split up into sections

which are transmitted piece by piece

over the radio transmitter back to Earth.

The photos look stunning,

but the quality of the image isn't very high.

The radio transmitter can send us the photos from a long way away,

but it's quite slow, so we can only send low-resolution pictures.

So, when I programme this, I have to choose what size image to take

and how much to compress it down to a file

that takes not too long to send.

So there's a compromise between a nice big image

that's got lots of detail but takes for ever to send

and then, a nice small image that we can send lots of.

All of this only works

because Dave's written some programs for the tiny computer

which make everything work together.

So what does the code look like?

It just looks like somebody's randomly done that.

Oh, yeah, that's how I build my programs.

It's not too bad actually.

One program tells the webcam to take the pictures

and what resolution to take them at.

This is written as a loop

so the webcam keeps taking pictures one after another, every 30 seconds.

Another program sends output to the radio transmitter,

transmitting the pictures to Earth.

It also sends the position from the sat nav tracker,

so that Dave and his friends can track the balloon.

With the balloon floating up above Earth,

Dave and his wife Julie set off to chase it.

The program on the minicomputer is constantly transmitting

its height and location,

so we know exactly where it is.

The balloon could land hundreds of kilometres away -

it's even landed in the sea before.

Let's hope that doesn't happen today.

Meanwhile, back at Mission Control,

with Dave's friend Anthony,

we view the pictures as well as tracking the balloon

and following Dave's progress.

OK, so those are just coming in now,

those are slightly higher than you would go in a plane.

If you just go up a little bit,

-you see the sky's beginning to get dark.

-Yeah.

-Yeah.

And you're getting more and more of the curvature of the Earth.

It's incredible, it's amazing.

I like that one because there's the sun,

the Earth and the black of space.

Yeah, it's beautiful, it's gorgeous.

That's one of my favourites because it's really bright.

Yeah. That's incredible.

It's amazing to think these photos were taken

by a computer the size of a credit card

floating above the clouds in a homemade box.

Dave even holds the record

for the highest photograph taken by amateur equipment,

nearly 40 kilometres above the Earth.

But today, our balloon is only travelling

up to about 30 kilometres.

That's still three times as high as a passenger plane flies.

Then, it bursts and starts its descent back to Earth by parachute.

-OK, there it goes, it's just burst.

-Oh, wow!

Looks like it's burst.

Oh, good, what altitude did it burst at?

30, 431.

30, 431. OK, that's good.

So Dave now knows, so he will go and track it down.

Yeah, he will have picked up in the car that it's burst

and he will now be going.

'After about 30 minutes, the balloon lands back on Earth.'

So that's the field it's landed in.

'Once the height we're receiving stops going down,

'we know the balloon's landed

'and, of course, we also get a picture of the ground.'

Can we talk to Dave?

Yeah, absolutely, yeah, if you type in that box there,

type in, "Dave, head up the M11."

We're heading left on the M11 in a few miles. Excellent!

There's a road on the left there,

so that's probably where we want Dave to get to.

As night falls, Dave pulls up in the field,

but will he be able to find the box in the dark?

There you go, so, how has it survived?

That looks pretty good, so almost completely intact,

and that's what's left of the balloon.

TELEPHONE RINGS

-Oh, there you go. Well, let's see where they're at.

-It's just moved.

Hiya, Dave.

Hey, Anthony, how are you doing?

-Hey, excellent.

-'I have some good news for you.'

I am holding in my hand

a parachute, the full piece of latex

and the Pi payload and buzz as well.

ALL: Yeah!

Superb!

'That's cool.'

They're quite pleased.

-Very pleased, yeah.

-Speak to you soon.

-'Thank you, bye-bye.'

-Bye, cheers, bye.

This has been a record-breaking piece of programming,

and we've been able to take some amazing pictures.

'By learning to program,

'there's all sorts of things we can make possible,

'whether it's creating a game for our mobile phone

'or doing something really cool with a tiny computer.'

The sky's the limit.

Today's computers are getting smaller and cheaper,

and this little thing is a Raspberry Pi.

It's a credit card-sized computer

and you can plug in your keyboard, mouse and monitor.

These have been designed to explore computer programming,

so I've come to St Saviour's Primary

to see how they're using their Pi

to generate code that has some real bite.

'Getting started is really straightforward.

'Just like any computer, plug in your inputs, such as your keyboard,

'and your outputs, in this case, your monitor.

'With the computers connected up,

'we're using open source coding software

'to programme a motion-sensing crocodile,

'so that when you place your finger in the croc's mouth,

'it'll bite you.'

It's now croc time, so you should see a box of Lego

along with some instructions - get going!

'So first off, following the instruction sheet,

'the group need to build their crocodile.'

Donna, where does this one go?

'The crocodile will then be brought to life by connecting it up

'to the miniature computer before we start the coding.'

-Is it those?

-Yeah!

OK, so what are all these bits?

So, this is the motor and this is the sensor,

so when the finger comes along,

-there will be a thing here and it will...

-Trap your finger.

Yeah, trap your finger.

So it will be the roof of the mouth.

And then, what does the motor do?

So the motor powers it to go up and down.

Up and down, oh, cool.

'So, in this system, the sensor is the input and the motor the output.'

So if you put your finger here, it will bite you?

Yeah.

So has everyone finished building their crocodiles?

ALL: Yes.

Awesome, now we're getting to the dangerous bit.

We want you to write a program, using Scratch,

so that when you put your finger in the crocodile's mouth,

it will bite you.

In the same way that the crocodile fits together piece by piece,

this computer code fits together like building blocks.

All the commands you need are on the left hand side of the screen,

which you drag into the middle to write your program.

For this program, we're going to need to use the sensing commands,

which tell the crocodile when to shut its mouth.

The sensor value tells you how far you are from the crocodile's mouth.

As your hand gets closer to the sensor, the number decreases,

so you need to check what number it goes down to

when your finger is in its mouth.

Now that the croc is programmed to know when to bite,

we need to programme it so it knows how to bite,

so we need to write some code for the croc's motor

so its mouth will open and shut.

Great, so you've programmed yours to start biting your finger, right?

-Yeah.

-OK, do you want to show me what you did?

So, as you can see, it's less than 20,

it won't bite unless it's under 20.

-So 20, that's the distance your finger is from the sensor.

-Yeah.

So when it closes, it stays on for one second,

and then, this way means it goes open,

so it goes that way for one second and then this way.

-So it bites your finger for one second and then lets go?

-Yeah.

-OK.

-And that's all down to your programming?

-Yeah.

Oh, that's so cool.

'And it really is that straightforward

'to write a simple computer program.'

Congratulations, you all built and coded a finger-biting crocodile.

Did anyone lose any fingers though?

ALL: No.

Good stuff.

'It's incredible to think you can do all this

'with such a tiny little computer.

'I wonder what else you can do with it?'

Robots aren't just in science-fiction stories,

they're already playing a role in our lives,

from making cars, toys and electronic gadgets -

we've even got them playing their own sport now.

WHISTLE

All robots are controlled by a computer,

but the question is, how does the computer tell them what to do?

My hackers and I are going to find out.

Hi, I'm Sam.

Hi, I'm Saffia.

We've come to Plymouth University to meet their robot football team.

Come on, guys!

One, two, three, go!

Today, it's people versus machines.

Sam and Saffia are using game controllers

to control these white robots,

but they're up against Plymouth's Black Ninja robots,

who play football in the International Robot World Cup.

The black robots have been programmed

so they can play by themselves,

they don't need anyone controlling them with joy pads.

It's pretty clever stuff.

These are our football robots.

The rules state, in the competition,

that all the robots have to be what's known as autonomous,

that means automatic.

So they have to do everything by themselves,

without humans controlling.

The main thing about these robots is that they use shapes

and colours to define things,

but they have to have little tricks,

like an orange thing on a green background.

If you took that orange ball

and you put it over in a different coloured background,

it won't think that's the ball.

You have to make very simple rules to make it understand what to do.

THEY CHEER

'To make robots play for themselves,

'we have to write computer code -

'it's this code that makes the robot do things,

'like walk or kick the ball.

'Programmers need to think hard

'about exactly what actions the robot needs to perform,

'what things does the program need to know

'and what things doesn't it need to worry about.

'For example, if we want to programme a robot goalkeeper

'we need to know where the ball is and when it's heading for the goal,

'so the position is the most important thing.

'We can think of a number line for our position,

'with zero in the middle where our goalkeeper starts.

'If the ball is heading for the negative numbers,

'then the goalkeeper dives to the right.

'If it's positive,

'then the goalkeeper dives to the left,

'and if it's near the middle, then he stays in the middle.

'But what does the code to control the goalkeeper actually look like?'

If the ball is between this far away and over to the left,

then we want you to dive left.

'If the position along the goal line

'is bigger than 100 but less than 2,000,

'then the code makes the goalkeeper dive to the left.

'But if the position is between minus 100 and minus 2,000,

'the goalkeeper dives right.'

And also, we have to have a robot that just stands in the middle,

because, otherwise, if he dived left

and the ball was going straight at us, it would be a goal.

'If the position is close to zero,

'in between minus 100 and positive 100,

'then the goalkeeper stays in the middle.

'So the goalkeeper decides what he's going to do

'using these "if" conditions.'

How do you write such a long and complicated code?

A long piece of code is made up of small functions,

and the functions do various little things that we want.

So "looking for ball" would be a function.

Then "walk to ball" would be another function.

So that makes it much more manageable.

So robots break down playing football

into a series of different functions, simple tasks.

I wonder how different that is from what Sam and Saffia are doing

when they're controlling their robots.

Oh, come on!

BOTH: Yeah!

Once a program has been written,

the code gets downloaded onto the robot.

Each robot has its own processor board,

a tiny computer that runs the code

so it can decide what the robot should do.

All right, guys, that's it, OK.

The scores are 2-0 to the white team.

THEY CHEER

'With so many things to do,

'playing football's actually a very complicated game

'for a robot to play.

'Even with all the long code,

'the autonomous robots couldn't always find the ball

'or kick it in the right direction,

'so they just couldn't win against Saffia and Sam.'

Press that button there and he'll do an impression of Usain Bolt.

HE CHUCKLES

It's a pretty good look, right?

So robots can be programmed to play football and other games

by breaking the tasks down into a series of simple instructions.

But this robot is called iCub,

and he doesn't need step by step instructions,

he can learn what to do for himself.

We looked at how children learn

and we tried to model that process

so that our robot is able to learn

in the same way that you or I can.

This is a box.

iCub: Box.

-Cool.

-What is this?

iCub: Box.

This is a stapler, what is this?

iCub: Stapler.

Where is the box?

'So iCub has been programmed in a different way.

'His computer program can learn new things,

'and that's a very complicated program to write.'

Does it only recognise one voice

or does it recognise other people's voices?

We use a commercial software for speech recognition,

and that's been trained on my voice.

Does it understand Northern, my accent?

-It would if we trained it on your voice, yes.

-OK.

Getting speech recognition to work with children is difficult

because the pitch is much higher.

'So the robot really only understands Tony's voice,

'but we're going to give it a go anyway.'

This is a box.

What is this?

'Try speaking deeper, try to sound more like Tony.'

This is a box.

What is this?

iCub: Box.

Yeah, that's cool!

This is a stapler, what is this?

iCub: Box.

Where is the box?

THEY CHEER

'Tony says that iCub's learning things like children do.

'I wonder how similar it is

'to the way people actually do learn things.'

Why did it get some things wrong?

If we'd written our program as a series of steps to execute,

then it would have gone through those steps

and it would have performed perfectly.

What's going on here with iCub,

because it's learning things in much the same way that children do,

it does make mistakes,

it makes quite a lot of mistakes

and that doesn't always make for the best demo.

What iCub's doing that's different is...

it's working in situations that the designer can't foresee -

you can't lay down a set of rules saying, you know,

"This is what a stapler is going to look like." All right?

Because then somebody will bring in some new stapler

that looks a bit different and iCub's lost.

'Learning a few objects might not seem like much,

'but this is early days.

'Programmers like Tony are constantly developing their programs

'and improving them so that one day in the future,

'robots will be able to learn to do all sorts of things.'

So there's still some way to go, but, like all technologies,

it only takes a matter of time.

Maybe in the future, you'll be working on the next generation

of robot football players.

Have you ever wanted to own your own robot?

Robots are really cool, but they're useless

if you can't tell them what you want them to do

in a way that they'll understand.

So I'm here at The Digital School House

with some pupils from Marish Primary School

and they're going to show me how they program their own robots.

We've got a map of a town,

full of buildings like hairdressers, florists and a gym.

Our car is going to start next to the sports shop at A

and then drive through the town to the baker's at B.

FD,

RT90,

FD.

First of all, we've got to plan our route,

breaking it down into a sequence of step-by-step instructions.

We can get around the whole town with just three instructions -

move forward, turn left 90 degrees and turn right 90 degrees, so that

our car follows the road through the town avoiding all buildings.

FD, FD, done.

Now, we're recreating the town on our computer,

using different pictures to represent the shops and buildings.

We're making this like a town so the car can move around.

We're trying to match these pictures with those pictures over there.

OK, so all these objects represent shops in the town, is that it?

Yeah.

So we go to gym, we press gym.

So you've just added gym to the objects of things

that now are on your map.

Yeah, so it will make it more interesting, cos then

the robotic car can go to lots of different places, like everyday life.

Oh, cool.

Now we create a procedure called "Drive".

A procedure is the list of instructions

that our car is going to follow. It will only do exactly what we tell it

so we've got to make sure we've got all the instructions right

otherwise the car will go the wrong way.

So can you tell me about the commands?

Well, I'm telling it to do forward, forward, right turn 90 degrees.

So the FD means forward and then the RT90 means right turn 90,

so that 90 is the degrees you want your car to turn.

These simple instructions are the building blocks of our code,

and with the right sequence of instructions

we can make the robot travel any route we like.

So, where will your car end up, then?

Next to the bakery.

-OK, fingers crossed, that's if everything works out.

-Yeah.

An important part of programming

is to test the code that we've written.

Some of the groups have decided to test their procedures

on the computer before they try it out with their actual robot.

That's a good idea, because it's going to be much quicker

to change it now rather than later.

Looks like the car's crashing into the gym.

You might need to check the code, guys!

Once we've tested it, we need to connect up our car

and download the procedure, so our code can tell the robot what to do.

And now, the moment of truth. Have we got it right?

We've made it past the first bend.

Hang on!

It's all gone wrong on this corner.

We should have turned right 90 degrees,

but we're going back the way we came. How did that happen?!

And now we've crashed into the cake shop.

That's what happens when you don't test your code properly!

Looking good so far.

So close!

But they've gone forward again.

They've missed the turn. Now they're not even on the map!

And, into the crash barrier.

When I was there, I was meant to do a RT90

then move forward twice and then do an LE90.

OK, so it was the right and left you got wrong, basically.

Oh, that's such a shame, guys!

-Do you think it's going to work?

-Yeah!

-Do you guys think it's going to work?

-Yes.

OK, let's go. Fingers crossed.

Tension's mounting!

Brilliant! We've made it!

That's great, guys! Next to the baker, just how we wanted to be.

Cool! It looks like it's going to be easy

just to write a set of instructions, doesn't it?

But actually, it's quite hard.

But you need those instructions to be precise, so that your robot

knows exactly where to go, that step-by-step process.

When you're programming, you often have to find and fix mistakes

until your code does exactly what you want.

We call mistakes "bugs",

and getting rid of the bugs is all part of the fun.

Everybody loves playing with toys,

whether it's train sets,

action figures or board games.

Many of today's toys already have computers in,

but there's a new generation of toys

which are way more sophisticated

making their way into toy stores near you.

To find out more about this,

my two intrepid hackers, Edward and Hannah, have come with me

to meet two guys who've created something quite extraordinary.

Hi, I'm Edward.

Hi, I'm Hannah.

They've produced something they call AppToys,

which use the newest technologies

and the most inventive computer programming.

So, guys, do you know what AppToys are?

-Well, I know what apps are.

-I know what toys are.

But we're not really sure what AppToys are.

OK, I think we're just about to find out.

Hi, guys. Come on down.

Thanks for coming. My name's Elliot, this is Martin.

Together, we're toy developers

and we'd like to show you our range of products.

AppToys use smart devices, such as phones or tablets, to control them.

First off, you download an app to your smart device,

which is then placed inside the toy,

such as a dog kennel or police car, which you buy from a toy shop.

When you put the phone inside the kennel

the phone recognises that it's inside the kennel.

Then it divides the screen up accordingly

so that it aligns with our mirrors, which gives it the 3D effect.

So, we have the phone that has the app we downloaded,

and then we've put it into this hardware, where there's mirrors,

but how does it all come together?

Are there sensors in the hardware, so the phone recognises it's there?

Yeah, there's accelerometers and the accelerometer recognises

what angle it's at, so if it's upside down,

which is what is required for the kennel, we can recognise that.

Smartphones are mini-computers

and have a number of different inputs, processors and outputs.

The accelerometer is one of these inputs

and tells the phone which way up it is.

The AppToy also uses the phone's other inputs, such as

the microphone, and processing software like speech recognition.

So, in this case, you give your virtual pets different commands.

Chase tail.

Wow!

-So it's using voice activation.

-That's good.

-Can it roll over?

-It can roll over. Roll over!

THEY LAUGH

-So how do you feed it?

-You say, "Feed",

then you put your finger through the front

and the camera is looking for your finger

so it knows that you've fed it.

Some of the toys make use of the smartphones' ability

to transfer information between them using Bluetooth technology.

This one's a bit scarier than a dog.

CREATURE ROARS

That's called Battle-Dino

because you train it to become a real fighting machine

and then you send it to your friend's crate

and they battle it out together, so the strongest dinosaur wins.

-Wow!

-That's really, really good.

That's cool?

-Good, we spent a long time working on it!

-HE LAUGHS

AppToys use the existing technologies of smartphones,

but they still need to program all this technology

to get it to do all the cool things they want it to.

Programming whizz Martin shows us how he used computer coding

to design the T-Rex game.

HIGH-PITCHED ROAR

In the game, the T-Rex will grow from a baby to an adult

and as he gets bigger, his roar will have to change.

Now, the really cool thing about coding and sound is that you can

use the code to change the sound, so you don't need to have lots of

different recordings of him roaring, you can just change it in code.

And what we did, if we go back and look at the T-Rex code, you can

see, when we go down here, we have this command, "audio.pitch", OK?

The pitch of the sound is the quality that makes it

either high or low, and what we're going to do is,

we just make that pitch equal to the speed that he's animating at.

The computer uses this number, the audio.pitch value,

to determine how long to play the sound for.

-So, can we have a go at that?

-Yeah, sure.

OK, so, instead of using the speed to do the pitch,

it could be anything, really, from 0 to 5, 6 or like 2.5, 3.5.

Shall we try 3.5?

-Yeah, go on.

-You type it.

3.5.

And then we put that to finish the line, then we'll save it.

So he's going to roar at 3.5 now.

SQUEAKY ROAR

THEY LAUGH

That, it wasn't scary, it sounded like he was in pain!

Like someone had stood on him!

SQUEAKY ROAR

Shall we make it a bit deeper?

Do you want to choose a number that will make it a bit deeper?

0.25.

Let's see what that sounds like.

BOOMING, ECHOING ROAR

If it gets too low, you won't even hear it, you'll just feel rumbling.

That was like a train going through a tunnel or something, wasn't it?

It was very deep.

BOOMING ROAR

And it's not just the pitch you need to code for.

The T-Rex needs a really precise set of instructions.

Commands to tell it when to grow, roar or when to walk.

And it's the same for all the games.

Computer code is written to control everything,

from the speed of the police car, to the bark of the dog.

DOG BARKS

It's been great to get a behind the scenes look at all these AppToys

and unlock some of the secrets of the coding.

But Elliot has got one last thing to show us - their amazing aquarium.

Oh, that's so cool, that looks so realistic!

-Can we go?

-Yeah.

-Oh, wow!

-It's epic.

-Wow.

-Looks 3D, right?

Yeah! It's looks like you've got real fish there.

So these things are actually there, and the rest is projected onto this?

-You've got it absolutely right.

-How does that work?

Well, we're reflecting the image from the tablet

onto some physical background that,

some of it you can see and some of it you can't see.

-So, can you see the doll on the right-hand side?

-Yeah.

So we've pre-programmed in that the fish know where that doll is,

so when they swim behind it we make them disappear,

so it gives the illusion the fish are actually there,

so that's why the whole thing looks so realistic.

It's also interactive, so I can fish in the fish tank.

-Whoa!

-Wow!

-That is so cool!

Wow, that's amazing! You can go fishing! Fishing in your home.

In this setup, the information is being transferred

between two devices.

The phone's accelerometer is the input

and the display from the tablet, the output.

I'm just trying to figure out how it's all working!

Well, it's using Bluetooth to connect between the two,

and I'm using the accelerometers in here

-to determine what I'm doing with it, so it knows whether I'm...

-FISHING REEL WHIRRS

Lifting it up or not.

-Go on, have a go.

-You go first.

FISHING REEL WHIRRS

-That's it.

-You have to wait for a fish to come to you.

Go on, pull up. Yes!

That was a big one, as well!

I think we'll give you the option to put it back in,

so that people don't get upset.

But this is only a prototype.

Yeah, well, what we're going to do is, when we launch these,

we're going to keep adding the features,

so it's never going to end, we're always going to put stuff in.

So if people come up with some good ideas, we can add it to it.

So, once we want to add a new feature

you just go back to the program, you write some new coding

-and then someone can download that app.

-Exactly.

I bet the makers of this smartphone never imagined

that it could be used to control a toy like this,

where a little bit of programming has allowed the creators

of these toys to use Bluetooth and speech recognition software

in such a creative way.

HIGH-PITCHED ROAR

So maybe in the future, teddy bears and dinosaurs will just appear

in your room, in 3D, when you open an app on your phone.

Subtitles by Red Bee Media Ltd