A New Type of Launch Vehicle: A Rocket with Artificial Intelligence

Yasuhiro Morita Project Manager, Epsilon Launch VehicleJAXA is developing a new type of solid–fuel rocket, called the Epsilon Launch Vehicle, aiming for launch in 2013. Japan’s solid–rocket technology is respected worldwide; the M–V launch vehicle, which was retired in 2006, had a reputation for the highest standards of performance. The Epsilon launch vehicle will reduce costs by incorporating artificial intelligence – a first in rocket history. A combination of existing and new technologies, this new type of rocket will have a great impact on space transportation.

Rocket Launch Control Via Desktop Computer

Q. Could you tell us about the next-generation solid-fuel rocket, the Epsilon launch vehicle?

Launch of the Epsilon Launch Vehicle (artist’s rendition)
Launch of the Epsilon Launch Vehicle (artist’s rendition)
Conventional launch control room (JAXA Uchinoura Space Center)
Conventional launch control room (JAXA Uchinoura Space Center)
Artist’s rendition of mobile control
Artist’s rendition of mobile control

The Epsilon launch vehicle is a three-stage solid-fuel rocket. It uses the existing H-IIA solid rocket booster as the first stage, and an upgraded version of the upper stage of the M-V launch vehicle as the second and third stages. Japan’s solid-rocket history began with horizontal flight tests of pencil rockets in 1955, and all the technologies accumulated over the last half century, from that time until the retirement of the M-V launch vehicle in 2006, are reflected in the Epsilon launch vehicle. Epsilon is the culmination of decades of advancement in Japanese solid-rocket technology, so I think its reliability and performance will be excellent.
However, Epsilon also uses new, cutting-edge technology. We aim to greatly simplify the launch system by using artificial intelligence. Today, a typical scenario is hundreds of people assembling at the launch center and working for several months in preparation for a launch. On the day of the launch, dozens of people are in the control room, monitoring every aspect. The Epsilon launch vehicle will drastically change this picture. You may doubt that artificial intelligence can be used in a rocket, but nowadays a self-inspection function is something commonly seen in machinery. Another example is a medical device such as the electrocardiograph, which uses artificial intelligence to diagnose heart abnormalities.
Ironically, rockets are seen as leading-edge technology, but the approach to their functionality and design is conservative. Rockets use technology from many generations ago, so they are like a showcase of deficiencies. There has long been a notion that new technology should be tested over an extended period of time before being used in actual launch vehicles. Consequently, the latest artificial intelligence applications have not yet been employed in rockets. The Epsilon launch vehicle will be the first rocket with artificial intelligence that will perform checks and monitor its own operation autonomously. This is the major difference from the conventional M-V launch vehicle. Q. What kind of new technologies are currently being developed? The most significant development in the Epsilon launch vehicle is the capability of running autonomous checks supported by artificial intelligence, which will enable rocket-launch control using a desktop computer (in reality, two computers for redundancy). We call this "mobile launch control."
The M-V launch vehicle required many devices for ground inspection prior to launch, and hence a lot of time and manpower. In addition, a number of components had to be manually assembled one-by-one at the launch center. As a result, from the time the first-stage rocket was positioned on the launch pad to the actual launch, it took about two months. With the Epsilon launch vehicle, we will be able to greatly reduce the volume of work and manpower because the rocket will perform checks autonomously. Furthermore, to reduce the number of rocket components, we are simplifying the rocket assembly so we can transfer the launch vehicle to the launch center in an almost fully assembled state. This will make it possible to launch the rocket only a week after its first stage is placed on the launch pad.
We are also working on reducing the weight of the propellant container, known as the motor case. The M-V rocket’s motor case was made of light and tough carbon fiber, and as a result it was the world’s lightest rocket in its class. Now, we are trying to make an even lighter motor case for the Epsilon launch vehicle, and also studying how to optimize the manufacturing process. In the conventional process of making a carbon fiber motor case, fiber saturated with resin, known as prepreg, is wrapped many times around a mold. Then, compressed under high pressure, the mold is fired at high heat for solidification. This is done in an autoclave, which supplies the high pressure, but that’s a large and expensive facility. For the Epsilon launch vehicle, we have improved the method of glue penetration so that the mold can be solidified without applying pressure. This has made the manufacturing process cheaper and simpler. Also, we have switched to a tougher carbon fiber and have been able to make a lighter and stronger motor case as a result. In this way, we obtain the higher performance structure by the lower cost.

Artificial Intelligence Helps Cut Costs

Q. How much cost reduction are you aiming for?

First small scientific satellite SPRINT-A
First small scientific satellite SPRINT-A

We plan to develop the Epsilon launch vehicle in two stages. The objective of the first stage is to launch a space telescope for planetary observation - the small scientific satellite SPRINT-A, scheduled for launch in 2013. The Epsilon launch vehicle will achieve drastic cost reductions with various innovations, including a shorter launch preparation time, mobile launch control, and an improved manufacturing process for the motor case. As a result, the estimated cost of the first launch in 2013 is about 3.8 billion yen - almost 50% less than the 7.5 billion yen price tag for the M-V launch vehicle. For low Earth orbit satellite launches, the launch capacity of the Epsilon rocket will be 1200 kilograms - about two thirds of the M-V launch vehicle, which had a capacity of 1800 kilograms. The cost will be reduced by half but at the expense of some of the launch capacity, so all told the cost performance will be improved by 25 percent.
Unfortunately, I think that cost is still relatively high to carry out more frequent launches. So the objective of the second stage of our development plan is to launch a rocket for under 3 billion yen by 2017. After the first launch in 2013, we hope to launch one rocket each year. This means that the launch in 2017 will probably be the fifth rocket. If we can drive down the cost of the Epsilon launch vehicle even further, monthly rocket launches, for example, will no longer be just a dream.

Q. Will the Epsilon launch vehicle carry only small satellites?

Our plan for now is to launch small satellites, with a total mass of 500 kilograms or less, but the Epsilon launch vehicle is capable of launching a low Earth orbit satellite with a mass of up to 1200 kilograms. A 1200 kilogram-class satellite is considered mid-sized.
The Epsilon rocket will also be able to launch a small planetary exploration spacecraft of around 300 kilograms. The asteroid explorer Hayabusa was 500 kilograms, so I think it is possible to devise a lighter, smaller exploration spacecraft. In fact, some researchers have shown an interest in using the Epsilon launch vehicle to launch spacecraft for lunar and planetary exploration. Small scientific satellites will be the main target for a while, but the Epsilon is full of potential. Although it’s only 24 meters high, shorter than the 30.8-meter M-V rocket, I believe it has greater potential than the M-V.

Miniaturization Is a Key to Space Development

Q. Does the Epsilon launch vehicle use the same propellant as the one used for the M-V rocket?

M-V rocket motor case
M-V rocket motor case
M-V launch vehicle
M-V launch vehicle

Yes, it is the same propellant. The first Epsilon rocket will take off in 2013, which is coming up soon. It takes so much time and labor to change the fuel that for the time being we will stick with propellant that’s been proven reliable. We are planning to develop new solid fuel in the next stage.
There is currently a wide variety of research on new fuel. One that we are studying is something that will fundamentally change the production method of solid-rocket fuel. Conventionally, solid rocket propellant is blended with a mixer in a half-liquid condition, poured into a motor case, then heated for shaping. Once it has solidified, it cannot be brought back to a half-liquid state, so there isn’t much time to conduct the entire process. And therefore, even if the required amount of propellant is prepared and divided into a few portions for mixing, it requires very large mixers. In addition, when transferring propellant into the motor case, you have to be very careful to ensure that there are absolutely no bubbles. It is a very intense one-shot deal.
On the other hand, the fuel we are developing is made completely the opposite way. This fuel will melt as many times as it is heated, and harden again at room temperature - just like a chocolate bar. That allows a fresh start when it is necessary. For these characteristics, it has the key benefit that the mixer can be small. Propellant is blended with a small mixer, produced little by little, nonstop, and can be stored like a chocolate bar. Then when there is enough, it can be melted and inserted into the launch vehicle so that it fits perfectly. This is a far more efficient method of fuelling a rocket. I think that one key to the future of space development is miniaturization, which increases efficiency.
A small mixer is also beneficial in terms of cost. A conventional large mixer not only costs billions of yen to make, but the maintenance fees are also immense, and on top of that, it involves many engineers, so personnel costs are high. With a small mixer, all these costs are reduced, so the cost of a rocket launch is a lot lower. Instead of just complaining over the shortage of funds for space development, I think that we need to come up with innovative ways to work within the limited budget we do have. Q. But if operations are streamlined too much, the volume of employment will be cut down as well. Won’t space manufacturers be hesitant about miniaturization? Certainly, some manufactures are reluctant to do so. However, without making machinery smaller and also reducing manpower, I think it will be difficult for the space industry to survive in the future. When a factory’s manufacturing process is streamlined, the profit from each unit may be reduced, but it won’t be a loss overall as long as production volume is increased. I think manufacturers and scientists have similar goals: they need to make a larger number of products to make a profit, and we want to launch many satellites.
Personally, I don’t dislike the current process of launch preparation, where many people gather and camp out at the launch center for a few months. In fact, that was one of the reasons I enjoyed my job. However, this style hasn’t changed at all since the Apollo era of the1960s. For the future of rocketry, in my opinion, we need to build a system where launch vehicles can take off frequently and come back like airplanes. Unless we change the existing approach and establish a system for rockets to be launched by just a few people, the next generation of rockets will never be born. Someone needs to pull the trigger at some point, and I believe it’s our job to do so.

  
1   2   3
Next