Thursday October 17th, 2002 Open in New Window
>> Good afternoon.
Welcome to the last session of the NASA on-line robotics course.
My name is Dave Lavery.
I'm in NASA headquarters in Washington DC.
Prior to the job I am fulfilling right now associated with the Mars exploration program I spent 12 years running NASA's research program in telerobotics technology.
We'll be talking about the things the last decade and a half and the next decade and what missions will we use, what robotics systems are like for space exploration and taking a look at where robots have operational uses as we try and expand our presence into the rest of the solar system and expand our knowledge of the rest of the universe.
Go ahead and put up the first chart.
Basically NASA's strategy for exploring the solar system, a couple different questions about what we are trying to learn.
We want to understand how the universe evolved, is there a possibility that life could exist somewhere else in the universe, and could we use the knowledge we gain by the solar system as we try and sense those questions to better understand what the Earth is like and how we exist in the context of the entire planet Earth.
And the fundamental knowledge required so we can actually establish a human presence in space long-term beyond the planet Earth.
Also ask how enabling technologies coming out of the process of answering these questions can make access to space and the aerospace environments cheaper and easier to do.
Finally, what technologies do we have to address, create that doesn't exist right now that will enable us to have the ability to answer these questions.
As we go through this strategy and try to define what the agency is doing, robots play a strong role in particular three of them, the question of life existing elsewhere, use robots to ser -- to search for the answer.
Improve quality of life on Earth.
As we explore the solar system, we can only do that with robotics system.
And how do we apply fundamental knowledge of what is required to establish a human presence throughout the rest of the solar system.
Robots will pave the way.
We can send them to the planets.
Right now, not feasible in the next couple of years to send humans there, at least without enormous expense.
They'll be the pathfinders to help us out at the planets.
Next chart, what we have done a number of years ago is take the strategy and try to figure out in the terms of robots fitting into the strategy wha, do we need to do, create, what sort of systems do we need to have developed that would allow us to have the right pieces, capabilities, functions, so we can answer those questions?
And what we did is we worked very hard to try to sort of distill it down to the best common denominator, figure out are there a couple of systems, if we could create the systems of robots with large capabilities, would they represent the components and bits and pieces we would need to build system to say answer the questions.
We developed originally about eight years ago, refined since and kept alive, and sort of the big target goals, the stretch goals we set out in the robotics research program to push the research, technology and our capabilities.
Five specific areas, I'll just quickly touch on each of the five.
You'll get an idea of the things we are trying to did.
Could we create a geologist, when a human geologist goes out in the field, they'll look around and get the context of the area, pick up a rock, study it, feel the texture of it, get a sense of the size of the grains on it, take a hammer, break it own, look at the interior, the difference between the inside and outside.
Sometimes smell rocks and taste them to get a sense of what the minerals might be, the chemical composition of the rock might be and the sensory information they use to form a -- we have a machine, so it could do maybe not all the job but a large part of it.
A rover out in the field by itself, onboard intelligence,Access a rock, look at the outside of it, and have enough onboard intelligence to understand what the appearance might tell you about the chemical or mineralogy cal context of the rock, maybe use a spetro graph, get a sense of what the rock is, make a theory, what is it, how did it get here, and send not just the raw data back to Earth, but what it is, and how it got there.
This is a build problem.
We know how to build a machine to drive across the field, but this is many years out in the future.
Establishing that as a goal is something we want to do early on.
Another one is large range rovers, they right now are typically limited to a few hundred meters.
The rover for 2003 has a total range of maybe a killometer.
What we want to do is have rovers that can travel not a few hundred meters or more, but travel on planetary scales, can they go thousands, and traverse from one side of Mars to the other.
And how you do that is not just limited to wheeled systems, maybe combination systems that drive and fly or drive and hop or go on ballistic trajectory, the goal was mobility.
Coordinated robots, build things in space.
We have a series of crews going up to the International space station, they are building it one piece at a time.
We are limited in terms of where people can go, where we can put human astronauts.
Are there robot systems, to go into a high polar orbit, the fundamental bits and pieces are there but we need to put it into a big system. The development of a space robot, a machine that has the equivalent as a human astronaut in a space suit.
The same level of dexterity.
If a human can use a tool in space, why can't a robot do it in space.
You have the capability, service a satellite built on the grounds by humans, designed to be serviced by astronauts in space, you can have the surrogate system doing the satellite instead of a human.
And when it's dangerous for a human outside, that's a preferable solution.
It's a big stretch goal, not there yet, but one of the things we are trying to achieve.
In the final one, it's the idea behind a lightweight, low cost, autonomous inspection on orbit.
You learned about the air cam, I have a model next to me, a full scale model.
The idea you could take the concept and create a fully autonomous vehicle so it didn't have to be maintained by an astronaut, I need you to go look at the thing I want you to look at and send back video, and then when you are done, come back.
Give it that level of commanding and have it act intelligently on its way.
So those are the big grand challenges we established a number of years ago.
I'm going to go through a series of missions and capabilities we have built to address some of the challenges.
Realistically right now, they are several years out.
Working towards them for a number of years and there is progress we have been able to make.
The next chart.
We have in front of you now a space robotics time line.
Talk about history and where we have been.
Robots in space are not new.
They have been around quite a while.
Go back in the earliest days in terms of what we call a robot in space, what I mean by that is a mechanism that can either be self-mobile, move itself around on the surface of the planet or in orbit or a ma -- manipulation on board.
The Surveyors had small arms on board.
Have small arms on board that can extend out, grab samples of soil from the ground, and it could pull them back on board and then analyze the soil.
Looking at the space robot's time line chart now, look across the top of the chart you'll see some of the earliest times that happened were in the late 1960s, 1967 was the first planetary space robot system flown aboard Surveyor.
The first time we went to the planets, the Surveyors went to the moon.
The first time on a planet was in 1976, Viking 1, 2 systems.
They, like the Surveyors, had a little manipulation arms on board, to reach out and grab soil samples from the surface of Mars and bring them onboard for analysis.
That was a step forward.
Up to that time we never had an ability to do anything other than look around at the cameras.
Actually be able to reach out and touch the environment, bring it back on board for analysis and understand what is there, react with the environment was a huge step forward.
Other things happened over the course of the years.
I think you are familiar with the idea of the RMS, a big arm on board the space shuttle, that was built by Canada.
Then a couple years not a whole lot happened.
There were not new systems being developed.
A lot were being flown but no new ones on board, until the middle of the 1990s and a big explosion of activity.
German flew an experiment, zero gravity, all of a sudden you had a free floating in space.
How did that differ from robot controlled standpoint?
We learned a lot since that.
The Mars Pathfinder mission with the Sojourner rover in 1997, a lot of activity, not just in the U.S.
Japanese have flown a mission, another manipulation they flew up on the space shuttle.
The Canadians have been developing a larger version of the space shuttle arm for the space station which was launched last year.
It has a special thing on it that is a dextrous hand at the end of the 50-foot long arm. The hand is a misnomer, not like a human hand, but it is nine feet across.
It allows them to pick up relatively small pay loads and do servicing of payloads on board the space station.
In the mid 1990s, a whole flurry of activity, new robots and new things being flown.
The next chart, the different realms of robots and how they get used in space.
Basically broken up in a couple different Regions.
I'll talk in three or four different classes.
The first is on-orbit systems.
Zero gravity. The next, a picture of a system called Charlotte, it is named after the spider from Charlotte's web, it's shaped like a box, from the corner of the box you extend cables out into the interior of the space spaition or the cabin.
By having the box with little motors attached to the cables, you could pull on the cables in different directions and you could move the box around in 3-dimensional space.
It allowed the system to do service anywhere inside.
In the space station, the cable, string it up to the corners of your work room and then Charlotte could move around by pulling itself around on the web around the work room, a gripper on the front to flip switches and maintain dials and pay loads.
The astronauts could do science rather than maintaining the payload.
Charlotte has been flown and we are looking at it, flown just on the space shuttle.
Looking at reuse and reflight on the space station.
Air cam, you heard about in an earlier lecture, a life-size, full-scale model, give you a sense of how big it is, one of the things unique about air cam, it's a robotic camera that uses these in the front as near and far sighted focus to give you different viewing length capabilities, free-flying.
It has a unique system that allows us to do at a very low cost and simple way to launch it as a complete spacecraft.
NASA takes this, tucks it under the arm, goes outside a space walk, holds out air cam and lets go.
At that point the spacecraft will fly around on its own.
When we need to recover it, fly it back to the astronaut, grab it, back inside.
It's an easy system to deploy and recover.
One of the unique aspects of it.
The next page, a picture of the space station mobile servicing system.
I mentioned a minute ago the Canadians had built a large space arm for the space station.
What you see in front of you is a picture of that system. The first part of it was just flown last year for the first time, and basically a large 50-foot long arm, sort of a big space crane that allows the astronaut crews to attach payloads, pick up large modules which may be tens of tons, and use this as a system to build the space station itself.
The picture, the multi jointed apparatus, the special purpose dextrous ma nape -- manipulation.
It's the hand of the servicing system.
It's hard to get a scale on that drawing, but what you see is the system is nine feet across from fingertip to finger trip.
Very large.
Considerably larger reach than the human astronaut would have.
What allows the astronaut to do is pick up large payloads, swap out modules, things like that.
Go to the next chart, see a system called Ranger.
Ranger was a project started about nine years ago and the idea was to take the concept behind the SPDM of a robot that can service modules in space.
But make it smaller, not so you have a nine-foot long reach, sort of a smaller, human's level scale.
You see an underwater version that's been developed up at the university of Maryland under the guidance of Doctor Dave acheen, what they have done, they have built a small version of the SPDM or a servicing robot to go in and attach to a payload, swap it out, a satellite free flying, for example, you could refuel it.
If a solar ray had failed to deploy, they could go in and repair the ray, power up the satellite, servicing tasks like that, but work at a scale a human astronaut would work at.
The picture you see there, the front of the Ranger robot, in the center of the image, it has a shoulder width span the same as a human, a work space about the same as a work space a human astronaut would have inside a suit, use the same sort of work volumes, the same sort of capabilities, could be controlled either from the ground or controlled by an astronaut inside a pressurized volume inside, but the own propulsion system, they could free fly around.
They could detach, fly around, come back in, do a servicing task to do a different work site over and over again.
That concept that you see the neutral boyancy version, we are looking to actually launch it and prove it in space. The next page, image entitled Exoskeleton systems.
Where Ranger looked at how a robot would interact with an environment, the other thing we want to look at is other than a mechanical gripper-type system, where you might see a conventional robot hand, two, three fingers that are large and look robotic when you see them, also want to look at the idea, can you make a robot not just in the dimensions of the shoulders and arms, but the hand, mimics a human hand.
So the robot system could actually use the same interfaces that a human would use.
If a human astronaut would use a pair of pliers or hammer or drill, why couldn't the robot use the very same tools and you wouldn't have to develop another set of tools.
Looking at Exoskeletons, mimic human hand motion in the image you have in front of you, that would allow a human to go through the motions of repairing a satellite or a maintenance task and a robot mimic exactly those motions and use his own tools to grab and pick up an object, maintain a satellite.
Those two ideas, if we go to the next page, you'll see two, Orbital Servicing Systems, those have progressed to the point where on the left-hand side of the page you see an image of Ranger in the free flying configuration and when we hope will happen in the future, actually have the human type or human form factor, human volume robot servicing system that can actually go up, free fly, move around, maintain satellites using the same interfaces and the ultimate realization is on the right-hand side of the page, the robonaut system, the robotic surrogate for an astronaut.
It uses the same interfaces on the robotic and tool side.
Mimics a human.
Whatever a human could do, the robot system is also capable of doing. The system you see in the photograph being involved at the Johnson space center and I actually had an opportunity to be down there in Houston just the day before yesterday, and what you see is a photograph about two months old and version two of robonaut is just being completed.
It is a case where you can sit and put on a glove and move your hands around, look to the side of you, robonaut will sit at the side of you and directly mimic the motions.
If I reach down and pick up the pen, robonaut will do exactly the same things.
Write, a very precise operation, just by moving my own natural motions, it will do the same things.
My ability to control the system using this Exoskeleton interface is very natural, fluid, and requires very, very little operator training or overhead to use the system and be productive with it.
That is a big step forward.
Let's go to the next chart in the next section, and which is planetary exploration technology.
Now looking at number 11.
These systems are fundamentally different than on-orbit Servicing Systems.
We are sitting down on the surface of planets.
What we want to do or have the robotic systems capable of exploring the surfaces, and in a gravity environment, the other big distinction between these and most of the on-orbit systems, they are far away, and a lot of time lag how we associate and communicate with them.
Most Servicing Systems maybe be the person operating the robot may be a few hundreds of meters away or worst case down on the surface of Earth and the communication link might be a few hundred miles.
Talking about planetary exploration systems, talking about a whole different realm.
The systems are being controlled by a human on Earth but the system itself may be millions, tens of millions, hundreds of millions miles away.
The delay time may be tens of minutes or in some cases tens of hours.
So you have to have a whole different level of technology, of intelligence on board the machines to allow them to be productive and understand the commands in a way that allows them to be productive for an entire day when you may be able to only talk to them once or twice a day.
The next page, page 12, a chart called scratching the surface of planetary exploration.
Basically what the chart is trying to show is for the last 40 years of history is the world, not just NASA, the world has gone out and tried to explore the planets.
One of the things sort of happened is we started on this trend where all the spacecraft we send now are getting closer and closer and closer to the surface and we are just now really the past couple of years starting to get to the point where we are getting machines on the surface of the planets and staying there for long periods of time and really getting a chance, rather than from orbit, just taking photographs and passing over a scene, actually able to get down on the surface, move around inside the scene, interact with the environment and see what's there.
And what's going to happen is over the next several years we'll be doing a lot more of this.
A couple missions that are mentioned on that chart, I won't go through all in detail, but get an idea of what's happening, around 2003, 2004, from that point on for the next decade there will be many missions flown by the United States, by the Europe people, and the Russians, Japanese, we'll have an intense intelligence to go out and conduct research on the planets.
What I'm going to walk through the next couple of pages is some examples of sort of what systems have led up to this point and then we'll talk about which ones we'll fly in the future.
Looking at right now, something called robie the rover.
The way he worked was it would sit there and use the big gold cabinet in the center of the machine, full of computers, the little gold bar across the top of the machine which housed four cameras, it would look at the scene in front of it and take the video signal, the computer inside the gold center chest would sit there and grind through the video signals for almost an hour.
>From that information it would understand what the environment was for about ten meters in front of it.
Computer aided remote driving, it would build a map what the scene was like, map out the hazards, a set of images, do a correlation to determine what the hazards were, where the rocks, terrain was in front of it, other obstacles, from that it would determine what a path was through the scene, and send back to the human operator this suggested path.
The human operator would say I like that path, go ahead and execute it.
Robie would drive forth a meter and stop and do it over again.
Sit there, one meter at a time, it would take about an hour to think through where it would want to go for the next meter.
A slow process.
We thought there has to be a better way to do this.
Started looking at other systems that could potentially increase the capability without requiring all the big computational overhead and about that time a student working at the jet propulsion lab for the summer developed this, tooth, he said there's a different way of doing things.
Rather than having to build systems and have cameras on board and understand, why don't we do what bugs do, wanders around.
If it bumps into it it senses when it bumps into it and backs up and reacts and finds a different way around.
Don't worry about planning things out, that's not what bugs do.
He built tooth, that's what tooth does.
It will wander around until it bumps into something.
It will back up, a very simple set of behaviors programmed in, back up, drive around to one side or the other and keep and trying until it finds its way around an obstacle, resumes the original heading and keeps on driving and doing its task.
A new way of looking at things that didn't require a lot of computational overhead.
It's a little hand wired computer that Colin built, the idea was one really intriguing, we ended up picking up on and following out into the future, this was an idea later flown on board the Sojourner rover, the same philosophy of driving until you encounter something and react when you get there, rather than preplanning all the motions.
The next page, you see a chart called volcanic fire walker.
The chart 14.
This is an image of a system called Dante.
A revolutionary idea.
A lot of people thought of rovers of wheels.
This was led by a university, if we wanted to get into intriguing places, harsh, wheels may not work.
Why not look at walking systems.
To test out the idea we built two of the robots, Dante 1, we pulled it out of the volcano, and we put 2 in mount spur, an active volcano in Alaska, we had the machine walk down inside, spend a week inside the volcano demonstrating walking machine technologies, showing how you could build planetary skill robots that could survive inside real harsh environments like the inside of a volcano and conduct experiments at the same time.
It took a science package on board and would walk over to an active gas jet inside the volcano and sit down next to it and for 24 hours at a time collect the gases.
Stand up and move around and do other investigations.
Spent a week doing this.
The first time we ever had a robot system anywhere in the world that had this sort of capability that was carrying a live science package, the interior of a volcano, a harsh environment, it embodied all the ideas of robot awe to no my up to that point.
You didn't have to tell it to lift a leg, put it down, lift a leg, put it down.
This is where I want you to go, you figure out how to get there, if you react to hazards, bump into them, keep going toward the target.
When you reach the target, notify my, I'll give you the next set of instructions.
Dante, we told it here is your target, we'll tell you how to get there.
You have to achieve it when you start talking about how to command planetary systems.
Have to be smart enough to be able to do several hours of activity on their own and just call back when A, either the task is completed, or B, in trouble and they need help.
Go to the next page, chart 15.
An example of Marscod.
Actually developed by the Russians and an idea they wanted to fly to Mars in 1996.
That did not include it, but we took a look at it as a neat idea how to build rovers for Mars, anything different than the United States was working on, we cooperated with them heavily, had them build several versions for us, we brought over, the folks at the NASA Ames research center took a look at how to combine it with the superb mechanical systems the Russians were working on and put them together.
It actually, a different size version is still in use as research vehicles and we are looking at them constantly to try to understand how you build intelligent robots that are capable of going out, navigating across different terrains, how the design of the mechanical system will affect how much intelligence you have to have on board.
For example, a lot more computer processing required to just move the legs of a Dante vehicle and have it understand where the legs need to be placed and look for good locations than you might have to have on board a wheel rover.
Conversely, maybe more rugged designs to get up and over rocks and able to avoid those hazards mechanically that a legged rover might not have to contend with.
It was really an exercise in trying to understand the mechanical advantages and limitations of wheeled vehicles, and how they can be combined with intelligent systems.
The next page.
Page 16.
This is another example of a research robot called nomad, also developed at the university by the way, that we sent again, looking for harsh environments to test robots, find places not ease ids for robots to survive in because if we can test them in harsh environments and understand where they do well and break, that will make us better prepared for the robots we send out to the planets.
This is a case we send nomad south, there was a meteroryte found, we know there have been many blown off planets and eventually come and land on Earth.
A collection of 23 of them we know originated on Mars and we suspect there are a lot more out there that haven't been found yet.
Nomad was built to wander around the antartic, and to find them.
Nomad was built by those folks, put it down in the antartic on two occasions to take traverses across the ice and use a sensor package on the front of it to scan for things on the ice.
When some of the large blue ice fields.
In addition, it has a unique job that no other robot is doing in the world, another unique thing is it's not controlled by a science team sitting next to it or the same continent.
They are in Pittsburgh, PITT university.
So from Pittsburgh, the team is able to communicate with the robot sitting on the surface of the ice in the antartic and drive it around, search for the additional meterorytes.
We say don't just drive it around, but use it to simulate how we would conduct science on the surface of Mars, simulate a Mars mission at the same time.
And then get a really good understanding of how you can conduct remote science.
Mentioned back at the beginning when a geologist sits there and walks around in the field looking for rocks, one thing he'll do is scan his head around, look around, find what is there using his own senses to get the context, then pick up a rock or two or three to start a study in detail.
It's hard in the eyes of a robot to get the idea quickly.
You can't take a quick scan.
You are usually limited, it might be an enter panorama, you have to do that in a different way.
Nomad gives us the opportunity to try to understand how you do science remotely through the eyes of a robot when you are not be there in person.
Try to exercise a lot of the different systems.
The next page, 17, the other things we are also looking at is not just the surface of planets but robotic systems under the surface, drilling systems.
One of the things we try to look at is are there capabilities or limitations behind how a robot would drill down in the surface, say, of Mars.
One of the things we know is intriguing about Mars is we think is that there are potential reservoirs of water locked up in the subsurface cht and we don't know if it's ten, 100 meters deep, if we could build a robotic drill of some sort to get down and access that, that would be important.
Number one, getting back to the exploration strategy, we are trying to search for signs of evidence of life.
Water in the subsurface, in all likelihood that's where the life would be as well.
The one thing we know is every case where we have any existence of life on the planet Earth, there is always water.
Water is a common factor.
If we are then led to believe that if there is life in the solar system, it's going to be where the water is.
If we can find water and we do know it exists elsewhere in the solar system, if there are signs of life is probably where the water is.
Let's go search for the water.
Getting the subsurface, drilling down with the robot-based systems is one of the ways we intend to do that.
Setting up a robot-based drill is not that simple of a problem.
Think about what you might want to do, a hand drill you might buy at the hardware store, all you want the drill to do was be able to drill a couple holes in a panel of plywood.
Think about all the things that happen when you pick up the drill in your own hand and drill through a piece of wood.
Target it accurately and put the point of the drill bit where you want to drill the hole.
You have to apply force in the right amount and resist the torque of the drill as it bites into the material.
Maintain a constant force, if you push it too heavily you can cause it to bind up and stall the drill.
All that has to be balanced out and all that is going on as you handle the drill by intuition and feel.
Well, we have to figure out ways the robot can do the same sort of thing, gather the same sort of information and be programmed and made intelligent enough so that it can do that autonomously while driving into the surface of Mars.
We know the surface will be covered with rocks, soil, things like that, and it has to be able to keep the drill clear, keep it from getting clogged, from getting jammed up and get to the surface.
You have an understanding of the sort of problems we'll run into.
Talking about the Mars mission, pick up the pace a little bit, page 19, a couple examples of what we are doing on Mars which is many of the robot systems are getting deployed.
What you have on 19 is a picture of the Sojourner rover on Mars Pathfinder, launched in 1997.
Page 20, see a couple examples of what Sojourner did and a couple pieces of information.
In addition to the science data the Sojourner brought back, also a lot of information about how robots in particular behave and what they can do in a Mars type of environment that we never had an opportunity to gather before.
Sojourner did a total of 234 commanded movements over the course of the 83 Mars days, slightly longer than an Earth day, so over the 234 commanded movements taking place during the 83 Mars days, we had a rover drive over 100 meters, so it didn't go that far.
Considering we never had a rover on a planet before at all, this was a big first.
Having a 100 meter traverse for the very first planetary traverse is not that bad of a deal.
Took over 500 images, a lot of soil experiments.
They were interesting, the way we did that, we wanted to find out how hard is the soil, dense manner, grainy, does it clump.
Sojourner has six wheels, we drive onto the soil, lock up five wheels and spin the sixth one and see how much energy it took to have it free spin and cut through the soil.
By watching the current required to make the motors turn we could back solve to figure out how much energy it took to make it cut through the soil.
>From that information we could then determine what the mechanical properties of the soil were like.
You got to the point where in addition where you could conventionally think of, things like the spectrometers, looking at the mechanical properties of the robot, understanding how they are working to infer scientific information about the environment.
Also returned over 245 megabits of data, that's probably not a lot if you are using an ethernet in the classroom, a couple seconds worth of data, but interplanetary data, it's more than we have had before.
So journ he were, -- Sojourner, all of a sudden rovers became a baseline to go on the missions that had never existed before.
It was the first.
Once we proved rovers could work on the surface of planets, they decided this is a feature they want to have on every future planetary surface mission.
Chart 21, the future of Mars exploration, I want to give a couple quick examples of some of the systems are going to be.
2003, launching a pair of rovers, big grown up cousins of Sojourner, the Mars exploration rovers.
In 2007, scout missions, the makeup is being decided right now, but could be small robotic rovers around the surface, multiple very tiny rovers, it could be a robotic aircraft.
One of the things we are looking as an opportunity.
2009 a larger rover called a Mars smarts lander, mobile science laboratory, that will go and traverse very long distances, all the information we gather from all the missions will be stacked up and from those once we have had the mission take place, we'll learn from it and figure out what are the big questions that then need to be answered in future missions beyond the turn of the decade.
And you see sort of a three-way fork in the road around 2009 that happens, basically tells us from what we learn in 03 and 05 and 07, make a decision around the turn of the decade what we want to do next, do we want to do deep subsurface missions, things like that.
It's hard to tell now until we find out more about Mars but at least we know what questions are likely to be asked.
The next page, 22.
Talks specifically about what's going to happen in 2003.
Basically as I mentioned, the big cousins of Sojourner in May and June, from Florida, they will live on the surface a while, be able to conduct a lot of science experiments. The other difference between the MER machines, Sojourner has the science things packed out on board, it couldn't reach out and touch with the soil, react with the soil unless it was with the wheels.
The MER will reach out, a little grinder on the front of it, grind away the weathered surface of the rock, get inside like a geologist would want to do, it has a microscopic things to look at it.
The MER rover will take a microscope, look at it, send back the images of the rock interior. The first time we have ever been able to do anything like this.
An exciting mission.
We'll be launching in May and June, landing in 2004.
Launching around 2009, the science lab, the big step forward is the fact these missions will be going much farther.
Rather than going -- these things may live one or two yards, very far as a potential.
The big thing, the big difference is there lifetime is increased, the mechanical system is more robust.
We can conduct three, four, five types of investigations at any one point with any set of instruments but we can do it over and over and over again across an entire region of Mars.
The next page, 24.
All that will culminate in a sample return.
They'll have an analysis, we want to pick up samples and bring them back to Earth.
And again in terms of a robot challenge, one of the hard things to do is get enough information, enough intelligence on board the machine to be able to understand the environment, look around, see what's there and have a robot pick out a good sample versus just any sample, and select that to be the one sent back to Earth for analysis. The missions are very expensive.
It may cost as much as one and a half to two billion dollars, if you are going to spend that much money you want to make sure the sample is a really, really good sample, and not just any sample you grabbed first.
We want to be able to select the best one, and work on the mechanisms, that once we have the sample selected, reach out, grab it, break it off, get the interior of the rock and bring it back.
Next page, 25, the last section we'll talk about, robots for the outer planets.
Basically this is a group of robot systems looking beyond the Mars environment to the rest of the solar system and seeing what is out there.
Go to 26.
An artist concept of another mission.
It is a slightly different type of mission.
It's a pure robotic spacecraft.
The unique thing is it would be able to land on a comet, rather than a planetary surface like Mars or a rocky planet or a larger gas planet, instead looking at a surface that's completely unknown.
We really don't know what the surface of comets are like.
It could be the consistency of cotton candy or hard concrete or granite or somewhere in between.
And we have to have a grapple system to attach it to the surface when you don't know how dense it is, and then the robot mechanism, it's a hard problem to try to solve.
Imagine trying to take a machine you want to build and have that breadth of potential surface service behind it.
Next page, 27.
System called the black smoker explorer which is, actually an artist concept of a remotely operated vehicle, underwater, doing some investigations around a black smoker on one of the Earth's ocean.
Trying to do when we build them, understand how to do science through the eyes of a robots.
Why is this a big deal at all?
We are looking at planets, and as far as we know there aren't any oceans sitting on the surface of Mars.
While that's true, you go to the next page, 28, you see an artist concept of something else, a pentraitor.
Lake bosstick is under a thick layer of ice.
On one of the moons of Jupiter there are ice crusts that may be very thick, but underneath the ice crust is a liquid water ocean.
If we can build robots to navigate around our oceans, do investigations, we are confident to go down under the ice, no direct human communication, that let us believe that we could send it to Europea, and explore under there, and send back signals about what it finds.
It is so far away you have to contend with delays, light time delays of an hour or more, and so the machine has to be intelligent enough to navigate around, swim around in the oceans and send back signals about what it's finding without relying us to drive around for it.
The next page, number 29.
Also looking at systems called nanorovers.
These are very small.
Robie weighed just over a ton, also other extremes over time.
Tooth was an example of what you can do instead of a large machine.
Tooth is in today's world big in comparison.
Look at things like nanorovers, this is about the size of say a cigar box.
Go to the next page, number 30, we want to push beyond that into what we call Robohands.
We want to get something the size of a cube of sugar or even smaller.
Why would you want to do something that small?
Well, the thing that's expensive about space flight missions is mass.
The heavier it is, the more expensive to get to the planets.
What we want to do is look at making very, very small machines that then are relatively inexpensive per machine to put down on the surface of Mars.
What that then gives me the option of doing is rather than flying a single large MER rover that may weigh 175 killograms and have one on each spacecraft or rocket, I have an option of saying I am fly one MER rover or I could fly 1,000 little robot ants, and I could throw these things out by the handful and each may not go very far, maybe a few feet, only investigate a very small region, answer one specific thing of information of a science question, but the whole total sum of information they send back is a lot and they can tell me about an entire region.
A different way of exploring and gives us a different option of how to conduct missions.
The next page, what is all this leading to?
We have the robotic systems out there.
What we want to do is learn about as much of the solar system as we can but the real reason is to answer the questions we have to in terms of understanding the planets, the surfaces and say is there a way, is there a reason and is it time to start looking at a human mission to the planets.
Just as we send out early explorers to the moon, early robotic systems before people, now to the planets and eventually those will be the precursors to the outer planets, to Mars, and when people get there, just as they did when we went to the moon, those robotic tools will be there to aid human exploration as well.
Images on page 31 in front of you right now, down at the bottom left-hand side of that page, you can see the idea of a human mission, artist's concept of a human mission on Mars, and what they have around them is the habitat, a number of different rovers that transport them around, transport their tools around, can assemble things for them while they are on the surface of Mars, in other words the robots are there to aid the people in the human exploration as well.
In addition to being a precursor, it goes along with people and does not in any way, shape or form, not does it not prevent the needs for humans but expands on the capabilities once they get there.
Page 32, eventually what we'll also see is the final culmination, the interesting part will be the humans will get to Mars and have an opportunity to use the robotic systems to explore around the surface of Mars.
One of the things we think will happen, sooner or later someone will find the first little robotic rover we sent to the surface of Mars perhaps a few decades before and find it and recover it and see how well it's done since it was in Mars.
Sojourner, we don't know precisely where it is.
We know where it landed, what it did while we were in communication with it, but what happened was when we lost communication with the Sojourner lander, with the Mars Pathfinder lander system, the rover basically communicated to Earth by relaying signals as it drove around the landing site, the batteries on board the lander died first, they gave out, before the rover died.
The rover was preprogrammed so when it lost contact with Earth, if it didn't hear from Earth it basically had a behavior programmed into it that said we think you may have gone perhaps a rock, you have lost sight of the lander.
Drive out and try and reestablish communication, that's what it thought was going on.
It tried to drive towards where it thought the lander was.
But we didn't want it to damage the lander, drive too close, if you get within ten meters of the lander don't go any closer and so the two behaviors would sort of battle each other in terms of what the rover was trying to do. The balance was the rover was driving a circle around where it thought the lander was.
And it would keep on doing this until it reestablished communication or eventually it died.
Well, over time its knowledge about where the lander was would tend to drift.
It was just using dead reckoning for navigation.
Over time, it would wander off toward the horizon.
Once the batteries died, Sojourner would keep driving in a circular path as long as the solar cells stayed alive.
We don't know how long it lasted, drove around in circles, but it kept on going for a while after the lander died.
One of the things we want to find out when the humans go to Mars and the Pathfinder landing site, where did Sojourner end up, how long did it last, survive on the surface, and when it got to wherever it went and had the systems give out, what was it trying to do?
With that, let's go to page 33.
I want to thank you all very much for your attention and participation in the course and open up for questions and answers.
>> Okay.
Sorry, I was very absorbed ip -- in what was going on.
We have lots of good questions here.
Where could we start?
There were a lot of good questions about the legs on rover.
Let's start with Christian's.
If a robot were to be built with legs, what would the optimum number of legs, the number of legs to have?
>> Good question, because actually that's, that's a topic that's always been hotly debated in the robotics community.
What really happens, it depends on the environment you are going into.
There are systems, you probably heard about them, that they are out now, some robotics from Honda, they walk in the way a human walks.
For a relatively benign environment, relatively flat surfaces they will do just fine.
A relatively benign environment, you make the argument, why not go with wheels.
The more rugged and harsh it becomes, you want legs to maintain stability and a foot hold.
The basic idea being you always need at least three good stable points in contact with the ground and those three points have to be outside your center of gravity.
Your center of graph if I, mass, has to be within the triangle defined by the three points.
As the robots moves you have to have the center of gravity within that triangle.
Enough legs to always guarantee you have at least three in contact with the ground, no matter what terrain is like and enough of them so as you move the center of gravity stays within that shape.
And some terrains the only way you can do that, as the rover moves forward, for example, and tries to take a step forward, you need to have four, five, six, maybe eight legs to guarantee the two conditions are met.
It's terrain driven is the real answer to the question.
In some cases two is all you need.
Some cases you may need eight.
>> Okay.
You mentioned wheels versus legs.
Brad is asking are robots with legs more efficient than robots with wheels, and Jenny would add, what kind of terrain on Mars would legs work better than wheels?
>> In terms of the efficiency, again it's all driven by the terrain you are driving across.
If you are on a nice flat, almost pool table, wheels will tend to be much more efficient. The rolling friction can be very low.
Very efficient and in relative terms they are very fast.
And so for that sort of environment wheels tend to be the white answer.
Conversely, if there is something that's basically a field strewn full of boulders and not a nice contiguous path through the field, then legs become advantageous.
Because you may have something with so many boulders in it the only way for a wheeled vehicle to get through the Boulder field is by driving up and over the Boulders, and as soon as they get to a certain size, about half the diameter of the wheel, it becomes very energy intensive.
If a legged machine, it can pick up a leg, go over the top of it, put it down, doesn't have to raise the center of gravity but just move the leg past that point, in relative terms, that's very energy efficient.
It depends again on the terrain, the number of obstacles, where they are, and how rugged it is, where the tradeoff takes place between wheels and legs.
A system developed many years ago, a machine could stand up to 16 feet tall, weighed a ton and a half, it could be powered, walk around, six-legged and walk around with less energy than a hair dryer.
But it was slow.
Take hours to cross a fair sized room.
It depends upon what the requirements are and what you are trying to do.
>> Okay.
One more about wheels.
I'm sorry, legs.
And this would be both Alex and Jacqueline are asking if a robot with legs stumbles or falls, how does it pick itself back up again?
>> That actually is a good question.
Hopefully the legged robot system will not stumble or fall in the first place.
That's a reason you might want to have redundant legs.
You may design a system that may need four or six legs, you may want to add two more for redundancy reasons, where if it stumbles, it doesn't tumble over.
There are ways you can bill a machine to self-right and stand back up, but all that adds mechanical complexity, and weight and mass, if you have to build a system to right itself, a huge potential power sync, so you want to make a machine that won't stumble or fall, and build it so if it does trip it won't fall over.
>> Franky wants to know, are these robots mainly for Mars or also used on Earth?
>> A lot of robots we are building, targeting for Mars exploration for the next several years.
One of the things we are looking at is how the systems can be used back on Earth.
Some of the early technologies that happened with robie and tooth and other ones, found their ways into things like robotic harvesting machines used in the agricultural community, we are looking at ways you can take, for example, demyeter, you can say go and cut the 50 acres of alfalfa without a human on board, and when you are done drive back and park.
We are able to traverse to a target and notify the driver when you are done.
Take the same technology, put it in the same application like cutting alfalfa fields, and it is not a concept, it's real.
The university and Ford new Holland company worked together to produce, and it's part of the product line you can get from new Holland.
>> That ties into another question, no, here we go.
Why robotics market has not yet arisen?
Will very so fis -- so sophistication, the robots work in the industry?
>> Two types of robotics markets.
At one level there is a good hell at this strong robotics, industrial community in place that goes back to the mid 1960s, the factory automation world, highly structured environments where you know the pieces you want to work with, assembly line, you know where the next piece is comeing from and it can be preprogrammed to pick up part A and attach to part B with full knowledge of where the stuff is.
That market has been in existence for a number of decades and is fairly healthy.
Most of the people who sell those machines, make and sell the machines tend to be either from Japan or Europe with a few Canadian companies.
There are not many U.S. industrial robots manufactures right now, a few but not many.
Other markets are field robots.
They work in unstructured environments, coal mines, agricultural environments, shipyards, why I don't know where things are precisely, I don't have a lot of knowledge about this piece will always be here where I need it.
Instead, in an alfalfa field there may be a rock over there, maybe not.
A hole here, maybe not.
I have to sense my environment to understand it and complete the task.
All the applications map directly to the planetary exploration.
This is a commercial market and robotability you want to buy.
It's just beginning to happen, just beginning to really start up in the past three, four, five years.
It's new.
I think over time it will eventually be much, much larger than the industrial automation market, factory robot systems, but just on the front edge of it.
Watch what happens over the next five years, I think you'll see things explode.
>> Frank wants to know, have you ever tried landing on a comet and studying it, and if not, there are plans to do that?
>> We have not yet had a mission to land on a comet.
It is in the plans for the future.
We have had fly-bys but not attempt to actually land on the surface of one.
One of the missions I mentioned was combine wd a project called DS4, going to be our first attempt to do a commentary landing.
Actually been pushed out a couple of years.
It is a plan, we have an intention of doing, but it hasn't been tried yet.
>> Okay.
Let's see, Daniel asks here, does it really matter what happened to the Sojourner from a scientific standpoint or just something the scientists would like to know?
>> Actually, from a scientific standpoint it doesn't matter that much. The science data it was going to tell us we got back.
But from a robotics standpoint it is interesting.
The engineering side rather than the scientific side of the question, we would like to know how long did it survive.
Sojourner was designed to just last seven days.
Only designed with a mission lifetime of ten meters traverse capability.
It lasted at least 83 days, drove over 100 meters instead of just ten, we know it actually surpassed by a large magnitude.
Did it last 20 times or 30 times longer than we thought it was supposed to?
And we want to know the answer to that question because that tells us an awful lot about how good our engineering is, and how robust the system is mechanically, electrically and other engineering aspects.
It's not a science question, it's an engineering question.
>> Jacqueline would like to know, what are some interesting things we have learned about our own planet while testing robots for space and streaming environments.
>> Some of the best examples of that question, the answer to that question are what we have been doing in the and the ar -- antartic and under the oceans.
We use them for what we think is under the ice on Europea.
We put them under the oceans, we interact with them like they are on the planets and we simulate missions.
The things we all do, while we are there, sort of rather than running a robot around and collecting the engineering information, why don't we take on some real science packages with us as long as we are there, and we collaborate with terrestrial scientists, geologist, they have questions to ask what the dark smokers under the ocean are like, we'll take their science packages with us on our robots and allow them to collect data.
They are learning more about the antartic, deep sea Regions, and biology questions are answered, what life forms may exist in the very extreme environments, the dry valleys, estimated up to a million years, no precipitation, how life can exist, we are beginning to understand as a result of these activities.
>> Okay.
We have two that really relate.
A little bit lengthy but bear with me here.
Nick asks NASA has used in past a strategy of cheap, quick, small rovers and satellites.
Are they going to continue the strategy or move toward large, more complex robots, and I think Eric asks a related question when he says if you make a lot of tiny robots, won't it end up costing about the same as one big one, while with the really hi-tech tiny microchips and the large number of the robots, I'm sorry, I didn't read that one through to get it clear, I hope it came across okay.
>> I think I have the gist of the question.
>> Great, okay.
>> In terms of the push towards multiple small cheap systems, I think what NASA has recognized, that was a concept known within the aerospace environment as the better, faster, cheaper concept, a push towards small missions that could be done in a relatively small period of time, one, two, three years instead of ten years.
And could return science that was a single set of science questions rather than trying to answer a multitude of science questions with the big missions that cost up to a billion dollars.
I think what we learned with the better, faster cheaper concept is it has a place and utility, and what it's done, we don't need to do them all as small, quick, unique or the large flag ship missions, now we have a choice.
We can pick which model of doing a mission works best for that particular set of investigations.
There are going to be certain situation where is there are large situations, the only way you can do it.
Large space telescopes, the only way they can work is you have to have to be able to collect a certain number of photons to work properly.
And so those sort of missions will be very large, single monolithic spacecraft.
The ant robots or nanorovers, we have the choice, the option that picks the one that works for us.
With regard to the question about does this idea of having the multiple little duplicate rovers, lots and lots of them, does that end up costing the same amount, well, it depends.
Again, yes, it can.
For the same amount of money I can have one large rover which may be appropriate for a particular investigation, or 1,000 little ones better suited for a different type of investigation. The model you might think of is each of the little rovers do one thing.
Preprogrammed with a simple sensor that says tell me if I find a certain type of rock here, basically a variant of salt only formed as best we know where there has been long-term standing water history, and so if you find that kind of rock you know that at one point in history there was a lot of water at that place, and that's, that may be all I want to know is where did the water used to be.
I take a thousand of these little rovers to ask the one question, is that rock I am sitting on, scatter them around the is your fashion of Mars and I fly over and say everyone who found that raise your hand.
It puts up a signal to the passing orbiter.
In one fell swoop I have 1,000 rovers waving in one place and I know that is where the rock is located.
That sort of model is what's attractive about the little nanorover concept.
It breaks easily.
Certain investigations it cannot do.
Little rovers won't go deep, pick up big chunks of samples, things like that.
That's where the larger rovers may become useful.
Cannot carry multiple instruments.
>> Okay.
Looks like we have a lot of interest in the little robot ant.
John wants to know how much one costs.
>> The one that I had in the picture, actually that was a research project started at MIT in 1995 or 1996 when we started that, and that was funded for several years at a total cost of a couple hundred thousand dollars.
If you want the first, it's pretty expensive.
Realistically now, the technology has progressed to the point, some people have just recently done more work in that area, now at the point where you can buy little tiny rovers like that for about $100 apiece, willing to put a little work into it and do some of the construction itself. The basic mechanism part is relatively cheap.
>> Okay.
Great.
Jenny wants to know, could an ant robot digit self down below the surface and dig there.
>> The problem becomes how do they communicate the knowledge, whatever they find, how do they communicate it back out.
The issue is they are fairly limited in terms of the energy on board, the batteries are small, the energy required to survive without coming back with a solar cell to recharge the batteries is limited.
In theory, yes.
Probably wouldn't be able to go very deep because it couldn't last very long otherwise, and has to have a way to get back up to the surface.
What you get is multiple dives subsurface but you will not go down meters.
>> Somebody must have anticipated where you were going with that.
Riley wants to know in connection with the ant robots, do we have the communication technology small enough to fit on them and enable to create a grouping of these robots?
>> To get a signal out, if you want to get a signal off one of the ant robots and back to Earth, if you are putting them, for example, on the surface of Mars at any meaningful rate is difficult, the communication platform is only a cubic inch, and because the physics associated with the radio signals says that to get the signal back you have to either have a very large antenna to radiate with or a lot of power, or you have to communicate at a low data rate, a minute or two.
It's that slow. The other option is find different ways to communicate other than radio waves.
One is have a orbiter in low Mars orbit that passes over periodically, and then it maybe has to only go a little ways, or use signals, the retrofleckor.
All it does is opens a mirror, waits for the orbiter to go by and paints the surface with a laser.
What the orbiter does, it looks for reflections from the laser back up to it and you flip the mirror open and closed and by Morse code, but the reality is, when you go small, the data rates are limited.
It will take you a while.
>> Chris wants to know what kinds of detectors are on the robots.
>> In the case of some larger systems, for example, the machines we'll be flying next year to Mars, the MER 2003, spectrometers, different electromagnetic radiation signals to try to understand what the chemical content of the rocks are.
We have something to look at the infrared bandwidth to look at the range, and matching that up against known spectra from rocks on Earth to determine, if this a variant of salt, something else.
And most of the instruments we typically carry on board, some of the large rovers like the MER rovers are different spectrometers, and each different bandwidth tells us something unique about the rock structure, chemical structure, oil structure, whether you are going to X-ray, whatever the instrument may be.
>> Okay.
There were several questions that came in while you were talking about the various types of robots.
Eric asks here, is Dante or maybe another one like that going to go to another planet or satellite?
>> Boy, is it ever going to go, that's sort of tough to answer.
At least for right now what we know is on the books for missions that will go to the planets, right now we don't have a legged machine that's planned.
However, the purpose of Dante when we built it was not necessarily to say legged machines are the way to go, but to show there are a lot of components technologies on board Dante, the vision, planning, mechanical motion system, the laser scanner used, all the different components were all brand-new and it was used to demonstrate and validate all those things, in addition to validateing the idea of walking machines.
What we have done, we have taken self -- taken several of the components and those have made their way to planetary missions.
Other things that actually happens and the way this happens is some of the students that worked on them, actually came to work for NASA.
Mark, who was working on the projects, the Dante and nomad systems, he took a system that he developed for them, and he is reuseing those algorithms.
That's how we see a lot of things getting reused.
Not that the specific reincarnation of it gets used but the components, technologies, and the bit technologies, and the people that do the machines bring the technologies that we create.
>> Couple questions about tooth.
Maria asks what's the point of tooth if it bumps into them and goes around them, and Alex says is tooth able to map out areas so it moves around so it won't keep bumping into the same thing constantly?
>> Tooth is interesting.
If you look at it with today's understanding of what robots are, it's a simple machine.
It doesn't do a whole lot.
It would bump into things, when it bumped into something based on which of the several bump sensors on board was activated it would back up and turn one direction or another and find its way around things, and feel its way around an object and continue on with the mission. The mission was to use the pair of pinchers in the fronts, find the hockey pucks and take them back to a beacon on the floor.
Simple task.
Tooth is ten years old.
Nobody had ever done that before with a machine this small.
A whole new way for designing control systems and architect things that no one had ever thought of.
It's a very simple task but the first time you do it it's important and difficult to solve.
It's only after we refined it for almost a decade that it now looks as simple as it is.
>> Okay.
Eric asks about Charlotte.
When will robots like Charlotte be able to operate without human supervision?
>> Actually Charlotte to a certain extent operated without human supervision while it was flown on the space Schultz.
McDonald Douglas was looking at, it gets back to commercial spinoffs, could you use something like Charlotte on Earth in addition to up in orbit and almost every application they come up with would be one wonder -- wonderful for an autonomous robot to do.
3 dimensions, you could use it as an application of a window washer that have sort of curved surfaces.
A wonderful application where you set up the robot, give it its work space, tell it what you wanted to do, a nice repetitive task on a large field, it can be programmed to be completely autonomous and act on its own throughout the entire task and go back to home base and say I'm finished or call for help when it gets in trouble.
Another application looked at was cleaning the sides of ships.
When large supertankers go in to be repainted and have the hulls cleaned, you have the side of a ship that can be several acres, a large blank surface.
Set up Charlotte, a curved surface, just having a simple, 2 dimensional machine will not work, set up Charlotte, take a sandblaster with it, a nice repetitive task, it can do the task on its own and calls for help if it needs it or call for operator intervention when its done.
Those sort of tasks are perfect for systems like Charlotte and those are things to be done today.
Not any need for further research on these.
>> Terrific.
You were talking a little bit about gathering of samples and I have two questions here, is there any danger of bringing dangerous unknown substances with the samples and how are samples protected on their way back to Earth.
>> Real good questions.
Actually a gentleman that works 70 feet down the hall from my office, his job is to protect Earth from contamination from interplanetary toxins and material as well as protect the other planets from contaminants from Earth on our missions.
They have defined a whole series of protocols and procedures we use, for example, every time we launch a spacecraft that's going to Mars, called the coast bar protocols, procedures used to determine if the spacecraft is clean of any Earth organisms that could be transport today Mars -- transported to Mars so we are not contaminating Mars with Earth bacteria or anything else.
Same community is specifying a bunch of procedures and standards, how we handle samples brought back to Earth, from Mars, asteroids and elsewhere from the solar system and how we are sure they are safe before they are brought down to the surface of Earth.
It's how the samples are packaged before they leave Mars, and once they are off the surface and on their way back to Earth, how the spacecraft is clean, and nothing comes down into Earth's atmosphere that's exposed, has ever been exposed to the Mars environment, a complete break of chain between things exposed on Mars and exposed on Earth, and you want to make sure that entire sample is tightly contained and safely protected, procedures we follow in terms of how the, a sample canister, once it's recovered on Earth i quarantined in a facility, like the CDC works with very dangerous Earth biology like the E bola virus, we have a similar place to guarantee they are safe before we do anything else with them.
>> Daniel asks or actually says, I am interesting in robotics field called beam, or BEAM.
Has there been any thought into using beam for your small ant robotics?
They don't need programming.
>> The beam robot, I think you are referring to the beam robot competition which has been around for a number of years is something we take a look @ its very core is something in terms of how to build robots with very simple purposes is a neat idea. The thing we found looking at it is that it's not something that works well in terms of specific and scientific investigations that you would want to pursue.
There are limits that we found in terms of how, when you make small robots, how dumb you can make them. The general philosophy, robie was a relatively smart robot, also big, a lot of computers had a lot of processing power required to make it work.
Went to tooth, much simpler.
It allowed us to do certain things but still relatively simple tasks.
When we found when we go to the nanarobots, we can make them smaller but also computationally smaller, dumber robots.
They may only be able to do one thing.
Beam is the ultimate I am me -- implementation, it is preprogrammed, an extremely simple task is fine, but ours are more complex than that so we have not pursued that too far.
>> We are drawing to the end of the hour and I do want to close with one question.
It is that you have talked about this being good preparation for a manned mission.
How far are we from that?
>> That's sort of the eternal question we always get asked.
The real answer right now is that we know somewhere out in the future there will be crude missions to the other planets.
It is in NASA's overall strategy, for example, that we are going to expand human presence beyond Earth orbit, we will be sending out human missions to Mars.
The thing we can't answer right now is when that is going to happen.
Technologically, could it happen in 15 years, yes.
That's a decision some other folks are going to have to ask.
It will be hugely expensive, very, very difficult and likely to be a mission so complex it's not going to be something that just one country will do.
Probably not something that just the United States will do.
It will be an International effort.
A lot of countries will have to commit the money and desire before we say it's going to occur.
It's going to but take a look and be very careful about sort of setting a date.
>> Okay.
I want to thank very much, thank you David for joining us here today, and filling us in on this really fascinating subject matter.
And I want to thank also the many students that sent in excellent questions.
This has been a great series, and I know the students doing this for credit will get e-mail ts, so you can do that, and thank you for joining us and we'll see you next time.
>> Thanks.