If you're thinking about an entrepreneur in the developing world, or

a conservationist, that wants to say, we've got this huge challenge when people

want air conditioning and refrigerators in the emerging economies,

not to mention light as well as processing for agriculture.

What is your advice to that entrepreneur?

Where the opportunities are for them to be able to benefit

from the huge challenges that are out there and turn them into opportunities.

>> Well, the thing that's been really exciting lately and particularly in this

off grid energy world is the rise of what we call remote monitoring.

So the challenge facing a lot of this energy access issue is that you might,

as an entrepreneur, go invest in a bunch of stand alone solar systems or

maybe small wind systems.

And you might go try sell those to people, individuals or communities,

that want to use them.

But those are by definition a little bit remote.

It's hard to provide the maintenance and support to that equipment in the field.

And it's also hard to find, particularly in developing context, people who have

the upfront capital to buy straight up the equipment you want to sell.

That challenge I think has held back entrepreneurs from being able to build

sustainable business models around providing this kind of equipment.

What we're seeing happening now is the rise of networked, monitored equipment.

Where the GSM,

often just 2G cell phone network, which is crazy slow by our standards, but

not bad in a lot of the world is ubiquitous and it's become totally viable.

And a number of companies have begun to use this business model to basically buy

a small cell phone antenna include that in the equipment you're selling,

the standalone solar or wind system.

Go back to your office and monitor daily

how that equipment is doing out in the field, even if it is 100 miles away.

You can get a small amount of data back from it on a regular basis to find out how

much did it produce, is it in good repair, how is it working?

And you're then able to do two things.

Number one, you can provide maintenance and support.

You can see, this equipment is actually starting to fail.

I can send a tech out now before it is totally offline.

And number two, you can do these very interesting pay-as-you-go

business models which are really starting to take off.

>> Could you explain what that is?

>> So this business model says, look, I know you can't afford this full stand

alone solar equipment, I'm not going to sell it to you, I'm instead going to

come install it your house and you may pay a small up front fee, but

then you're going to basically pay a subscription.

You're going to use, often, mobile money to pay me,

top up your account, for every kilowatt hour that you use from this system.

And I'm going to know how much you use,

because I'm monitoring you, I'm monitoring your equipment remotely.

So every week or month, you go on your mobile money account,

you top it up with some relatively small amount of money, and the lights stay on.

And this means you didn't have to try to save up a big chunk of up front

capital to buy the equipment.

In a sense, I'm financing you but I am building a sustainable business model

because I am getting revenue in from all of my customers.

And I'm tracking them, I've built sort of, in a sense, a wireless or

cellular electricity meter without having to run any wires or

install any very expensive upfront meter.

So to my mind that is an extremely interesting business model and

it is absolutely taking off.

That's a place to really really get into.

>> So if you had $1 million, let's say that we were talking, would we see

McCormick Utilities replacing Pepco in Nigeria?

Or- >> Well, maybe.

I think, it's possible, either in the electricity space, or

increasing also in the water space.

This kind of remote monitoring is a big deal for wells and other water provision.

>> Why does that matter?

I can understand why it matters for electricity.

How does it matter for water?

>> Largely it's this maintenance and repair question, so

you might install a well and then it may not function or break.

Being able to provide ongoing water services in a way that's not

guaranteed but is likely to be sustained because someone is monitoring it remotely,

that's a much better proposition for everybody concerned.

>> But you could also price water, I meant.

>> And you could also price water >> [CROSSTALK] And based on scarcity,

climate change, and other things that might be.

Because I think you're from the west.

>> Yeah, you can do that.

Now you know that this is a difficult thing to start charging for water.

We see what's happening in Detroit on charging water.

Actually, one interesting point I think it's important to remember on all this

distributed stand alone stuff also is that even in cases where there's plenty of

water, where it's not a fundamental scarcity problem,

mismanagement by either large utility companies, or government, or

both, can often lead to huge water shortages.

So the big example right now is Sao Paulo in Brazil,

where there is no shortage of water in the country.

It's the Saudi Arabia of water.

Sao Paulo is in deep, deep trouble.

Taps are actually starting to run dry near the end of the day.

And, it really looks to everything I see, as if

what was happening is a mismanagement problem, not a scarcity problem.

And the idea of sort of standalone distributed electricity and

water, in part, is that you, the user, the owner, begins to get to the point where

you control the production and consumption of that water.

You're not reliant on a large utility or

a government which may or may not be reliable.

So I think that's a key power- >> Does it actually make the system more

resilient?

>> Absolutely.

So that means if there is a power failure,

we saw actually in India two years ago now, I think, the largest

blackout in the world, 600 million people affected by, for a day with power failure.

Because the grid was being run unreliably and

the grid operators were not totally aware of where power flows were.

Failure in part of the grid took out an enormous fraction of

humanity's electricity.

If those people had had stand alone rooftop solar systems,

we wouldn't see that at all.

>> Or, if they had the data to actually monitor those systems.

>> That's the other approach, absolutely.

So we can't abandon the idea of running the grid well, totally,

we need to make sure that's being done well as well.

>> Getting the data, do you see data as infrastructure as much as utilities or

anything else are for countries nowadays.

>> Absolutely, I mean, what we could do with data now is extraordinary.

I mean, we are not exploiting that well.

We're learning slowly about how to use it.

>> Okay, how should we exploit it?

>> Well, so let's see, well for electricity and other utility systems,

we should be using much better sensors.

We should understand how to monitor power flows, water flows.

That's becoming cheaper and cheaper.

I often hear from government policy types, that would be great but

I can't afford a billion dollar sensor network across my entire system.

That's not what we're talking about.

We're talking about a much, much cheaper investment if it's done right.

That's for the sort of idea of In-situ Sensors, water level sensors, or

electric power flow sensors.

They can use data, transmission that rides on cell phone networks and

doesn't require big wires to run everywhere.

Another place the data, I think is boom and

is fascinating, is the satellite imagery world.

Where for years,

satellite imagery is something that's been available commercial, also of

course available in the intelligence world but we'll put that aside.

The commercially available satellite imagery has been extremely expensive

because the business model is that a satellite company will launch a small

number, two or three billion dollar satellites with very high resolution

cameras, and will then take images according to commercial requests,

very high quality, and sell them for really quite high amounts of money.

What we are now seeing in the last few years is an absolute revolution in

the way that satellite imagery is being done.

This is driven, I think, by the rise of these nano satellites, these tiny, cheap

and dirty and quick, hold in your hand little satellites that are being built for

just a fraction of the cost of what these high end billion dollar satellites are.

Something in the order of 50k maybe.

Maybe.

They're fundamentally enabled by this hardware revolution, a lot of

the electronics in them is the same stuff you can buy for $25 at Radio Shack.

The cameras on them that are taking images,

which are not as good as the billion dollar ones but

are extraordinarily much cheaper off the shelf, high end but off the shelf cameras.

So, startups, one that's spinning out of NASA and then other ones that are popping

up all around the world are beginning to build and launch these in high numbers.

>> And you're working with one.

>> And I'm working with one, absolutely.

And we're seeing the creation of value chain, where these satellite

computers are going to be operating networks of dozens or hundreds of these

little satellites that can take imagery all over the planet frequently, and

are finding ways now to then make that imagery available, not for charity,

but at much lower cost than what previously would be available.

That kind of imagery can do enormous things for managing water supplies.

You can see reservoir levels and lake levels across huge amounts of territory,

you can look at road networks and understand where there might be

landslides, where there might be traffic conditions.

All kinds of other sensing that you can do.

The challenge is going to be having large scale utilities and governments,

that haven't thought about this stuff at all, realize that opportunity and

start to embrace it.

And to do that, they're going to need entrepreneurs and

they're going to need technical people to inform them about where

the technology is now, where it's going, and what the opportunities are.

>> And we've actually seen a high school in Virginia that launched

its own- >> Absolutely,

Thomas Jefferson High School, you bet.

>> So this is something that- >> This is in reach.

>> This is in reach for almost anyone >> One more thought, sorry.

Just to come back to the developing world on that, we are also seeing these nano

satellites being build and launched by developing countries.

Nigeria, Columbia,

these countries are able to begin investing in

a space program, or an actual usage of data systems from space,

at a price point and a cost level that is totally reasonable.

This is not NASA investment stuff, this is stuff that can be done by smart,

talented people without a whole lot of resources in developing countries.

And I think that's extremely exciting and interesting and

is going to lead to all kinds of usage of that data.

I've got a whole bunch of little sensors up here we're going to talk about and

I'm going to try to tell you what the heck sensors in the hardware

evolution actually mean.

So here's a quick outline.

What are they? How do you use them?

And some example projects to show you what they mean and what they can do.

So what are they?

Well basically sensors are things that measure the world.

They measure things like temperature, pressure, humidity, location,

whether a door is closed, whether somebody is looking at you,

whatever you're interested in knowing, a sensor is something that measures that.

Another term we don't hear enough but

should be coupled intimately with sensors is actuators.

Fancy, fancy word, it basically means stuff that does stuff in the world.

A motor, a pump, a fan, a light,

a screen, it's anything that will then make something happen in the real world.

And the key piece in the middle of this is what's called a microcontroller.

A microcontroller is sort of like a miniature computer and

it is the thing that, basically, listens to sensors to learn about

what's going on in the world, and then tells actuators to go do something.

So it's sort of the brains in the middle of that equation.

Here's a quick run down of some sensors.

So, I've got a whole pile of them up here.

Just to give you a sense of scale, a lot of the sensors that are now available on

the market for a few dollars each, are like this.

They're sort of the scale of a postage stamp,

you can buy them in really cheaply, a whole pile of them here.

Some examples are digital thermometers, something called

a passive infrared proximity sensor, it tells you, it pings if somebody's nearby.

A humidity sensors,

a proximity sensor which works by a different principle, an accelerometer,

if it's being jerked around in car or something, it'll say, it'll tell you that.

Ambient light, so is it bright or dark, and how bright or dark is it?

GPS location, ultrasonic distance, how far away is the thing that I'm pointed at?

Gas sensor down on the bottom, which can tell you, is there smoke,

methane, carbon monoxide, it'll ping off if there's any of that stuff.

All those are out there and more, many, many more.

I didn't have space on the slide.

Mostly, a few bucks each.

So how do you use them?

Real quick and dirty.

Couple of sensors rigged up to a microcontroller, and

then that microcontroller in the most basic way you want to do it,

can just sort of display on a screen or tell you here's what I'm measuring.

It's just a digital thermometer.

Here's how warm it is, or here's how humid it's been, or here's how bright it's been

in this room for the past 48 hours, boom, boom, boom, boom, boom.

Microcontroller.

What's amazing, in the last few years, is you used to have to sort of think about

this stuff with full on computers, if you wanted to put a sensor system together

where you measure the humidity in a room and then do something with it,

you kind of have to dedicate a whole computer to it.

And that's what a lot of industrial control in factories has used,

is computers for it.

What we've seen lately, in the last few years, is a revolution in these guys,

in microcontrollers.

The availability for this is $25.

Availability of basically miniature computers that are programmable and

that you can connect to a bunch of sensors.

They can read those sensors.

They can then either, spit that information,

just tell you, hey, I've been tracking smog levels in downtown, and

here is what they are, every hour for the last seven days.

Or, they can also take that data and act on it, and do something to a motor,

a pump, a fan, or whatever it is that you want to do.

So, let me get into that in a minute, but just give you a run down of kind of the,

the zoology, of some of these things.

The sort of precursor to all these was the Arduino, this little guy, comes in six or

eight different varieties now.

The tiniest little one which is just itty bitty is down at sort of 10 or

15 bucks, just incredibly cheap.

The bigger ones that are slightly more powerful might be twice the size of this

thing, two or three credit cards sizes and get up to 75,

but you've got a whole range in between there.

Raspberry Pi, also fantastic, down at the $35 range, also about the same size.

BeagleBone, Intel Edison, Intel has gotten into the game here and

totally convinced that we're going to see a bunch more of these show ups.

So, little guys, cheap and programmable.

So, this is the only math in the talk, very obvious math I hope, low-cost sensors

plus low-cost microcontrollers is a cheap basic sensing system.

So then, let's talk about the actuators side.

Stuff to do stuff in the world.

LED's blink, blink, blink.

Motors that will turn things.

Pumps.

TV screens.

Fans.

All this kind of stuff that you might want to use.

And let me just give you an illustrative example, here we go, how do you use them,

here's sort of how you conceptually attach them.

But let's talk real quick about a greenhouse example.

So, you've got some awesome plants, they're in kind of a greenhouse

in your house, you want to make sure they get watered enough but you're going away

for a week and you're not really sure that you'll be able to make sure that they're

watered, totally possible to buy, pretty cheaply, soil moisture sensors.

So kind of a little thing like this that sticks down in dirt and

tells you how wet that dirt is.

And an ambient light sensor.

You could hook those two things up to an Arduino, put the ambient light sensor in

the window, put the soil moisture sensor down into the dirt that the plants are in.

Then hooked it up to arduino will read every, you set it up for, let's say,

every 15 minutes to read how, is the sun out?

Is it bright?

And is the soil still moist or is it dried out?

Then you hook the other side of that up to an automatic little watering system.

So you've got a little bit of water and a little pump.

And the Arduino when it decides, it's getting dry and it's pretty bright out,

I'm going to turn the pump on for the next five minutes.

That then automatically irrigates, waters your plants, and

when the system is set up or programmed it to stop, it will turn the pump off and

you just automatically made sure that, even though you didn't touch a thing,

your plants stayed watered in response to just how wet or dry they were.

So that's sort of a rough and dirty how you would do a sensors and

actuaries system in combination.

And the total cost of that kind of thing is 50 bucks or less.

Where can I start?

Here's a whole bunch of really fun resources around sensors and hardware.

So everything from say, Make magazine.

There's some getting started with Arduino and Raspberry Pi, companies like Sparkfun,

Adafruit.

Arduino hack-a-day is a great place, a bunch of stuff at Radio Shack.

Instructables is a great place, it has all kinds of step one, step two,

step three ways to build stuff.

Public lab did that smart phone spectrometer.

And a ton more.

So a whole bunch of very cool places to get resource to figure out how to use,

how to get into this sort of sensors and microcontroller thing and

how to do interesting useful projects for them.

So again my questions, now you know about that, what will you make?

Thanks.

Sensors

Innovation and Design for Global Grand Challenges

Rencontrer les enseignants