MIPS in space: Inside JAXA’s Hayabusa-2 mission to asteroid rendezvous
Over three weeks ago, Imagination Technologies had an exclusive interview with the New Horizons engineering team. The high-profile story presented the hardware and software systems that took the probe to Pluto and beyond.
A week later, Imagination Technologies revisited another important space mission called Hayabusa-2. Back in July, Imagination Technologies wrote about the 64-bit MIPS CPU aboard the Japan Aerospace Exploration Agency (JAXA) spacecraft headed for the newly-named 162173 Ryugu asteroid. After a brief exchange about the chip and how it had been designed, Imagination Technologies decided to organize a full interview with the engineering team working on Hayabusa-2 to find out more about the probe and its mission.
Rather than retell their story, Spaceflight Insider has opted to have Alexandru Voica of Imagination Technologies print this interview in its entirety. His contact at JAXA and the person replying to the questions below is Associate Professor Yuichi Tsuda, Ph.D., the project manager for Hayabusa-2. Dr. Tsuda works for the Department of Space Flight Systems at the Institute of Space and Astronautical Science (ISAS) and at JAXA.
A short Q&A with the JAXA Hayabusa-2 team
- Please provide a bit of your background and your involvement with the JAXA Hayabusa-2 project.
I was assigned the position of project manager for Hayabusa-2 right after the launch [Ed: April 2015]. Until then, I was part of the engineering team who worked on Hayabusa-2 from its initial development until the launch. I was responsible for designing the spacecraft design and tracking its development as well as other engineering aspects related to the project.
- Can you offer a brief summary of the scientific and engineering goals of the Hayabusa-2 asteroid explorer?
Hayabusa-2 would be the world second round-trip mission to the asteroid, following Hayabusa-1, and the world first round-trip mission to a C-type asteroid.
The main goal of the Hayabusa-2 mission is to visit the near-Earth asteroid 1999 JU3, conduct on-site science experiments, collect soil samples from the asteroid and return them back to Earth. 1999 JU3 is a near-Earth C-type asteroid and is believed to contain organic and hydrated minerals. We expect these samples to provide fundamental information regarding the origin and evolution of terrestrial planets as well as the origin of water and organic life on Earth.
The Hayabusa-2 mission has other goals, each split in scientific, engineering, and exploratory categories. The first scientific objective is to observe a C-type asteroid at various scales and investigate the interactions between mineral, water, and organic compounds. This should reveal sub-surface materials and formation mechanisms of the asteroid; we plan to expose these materials by drilling artificial craters into the surface of Ryugu. The probe will also observe the real-life conditions and dynamics of crater generation in space and its reaccumulation process.
On the engineering side, we plan to increase the robustness, reliability, and operationality of the sample return technology demonstrated on Hayabusa-1. In addition, our engineering team will generate artificial craters by kinetic impact which is a unique opportunity to demonstrate this technology.
The last aspect – exploration – is an outcome of the above two objectives. Nowadays, space exploration using small robots is considered to be one of humanity’s most important way to expand its activity in outer space. Hayabusa-2 is positioned as JAXA’s strategic mission to perform and expand on this activity.
- The original Hayabusa mission used a 32-bit CPU while Hayabusa-2 was upgraded to a 64-bit MIPS-based chip called HIREC HR5000. Can you explain to our readers why you’ve switched to a 64-bit CPU for this space mission? What do you like about the MIPS64 architecture?
This RTOS is well supported on the MIPS64 architecture and by this microprocessor in particular. NEC, the manufacturer of the spacecraft, has libraries for automatic and autonomous operation of satellites developed on TRON-based RTOS.
- Presently, mobile chips are manufactured using conventional bulk CMOS technologies while processors used in space missions employ SOI. Can you explain the difference and why SOI technologies are better for space travel?
The design strategies used for SOI in order to prevent the so-called soft errors caused by cosmic radiation are more straightforward than those of bulk CMOS. A simple design based on this strategy is preferred for high reliability in space applications.
- The first Hayabusa mission was an amazing success. What were the lessons learned and what did you improve in terms of hardware and software design for Hayabusa-2?
The hardware inside Hayabusa-2 has been fully upgraded using state-of-the-art components. Very briefly, Hayabusa-2 is heavier than Hayabusa-1 by 100 kg. Fifty percent of the additional mass is allocated for increasing redundancy of each subsystem, and the rest 50 percent is used to increase the scientific capabilities of the probe.
In terms of software, there were no anomalies regarding the data handling system on Hayabusa-1. Therefore, we tried to keep the same design process for Hayabusa-2 even though the development time scale was almost half compared with the prior mission.
The experience of Hayabusa-1 proved very valuable in shortening the development schedule while keeping the same reliability control process.
- The new Boeing 787 Dreamliner employs about 7 million lines of mission critical code. Can you give us a feel for the code complexity in the case of Hayabusa-2? What are the performance and power requirements that MIPS CPUs in space need to handle?
The length (complexity) of the Hayabusa-1 and Hayabusa-2 mission critical code is around a million lines. A lot of effort was paid to minimize memory requirements, and to improve the code size. Thanks to the MIPS architecture, we were also able to keep the power consumption of the microprocessor for Hayabusa-2 the same as that of Hayabusa-1, but with far higher performance.
- You are currently running a campaign to name the (162173) 1999 JU3 asteroid that Hayabusa-2 has been sent to visit. What are some of the more interesting suggestions you’ve received so far?
We received many interesting suggestions. I cannot give you specific examples, but we’ve selected a very exciting name, both for us team members and for people all over the world. [Ed: Read more about the meaning behind the name here.]
- Finally, can you tell us if you are planning a Hayabusa-3 mission? Are there any other MIPS-powered probes in the works?
We have plans to explore the Moon, Mars, and Jupiter in the first half of the 2020s. Since at JAXA we’ve adopted a MIPS-based chip as our default microprocessor, I think our near-future missions will be designed around MIPS-powered probes.
ABOUT THE AUTHOR
After having experienced the fast-paced world of the IP business as a junior engineer at various companies around Europe, Alexandru Voica has decided to pursue his dream of working in technology marketing and PR for Imagination Technologies.
His background includes research in computer graphics at the School of Advanced Studies Sant’Anna in Pisa and a brief stint as a CPU engineer. When not planted firmly in front of his laptop, Alexandru can be found hitting the basketball court, singing along at a rock n’ roll concert, enjoying art cinema or reading his favorite American authors.
Voica reached out to SpaceFlight Insider to share his experiences with our readers through content he had produced earlier. The original version of his article can be viewed here: Hayabusa-2
You can follow him on Twitter @alexvoica. View all posts by Alexandru Voica
SpaceFlight Insider is a space journal working to break the pattern of bias prevalent among other media outlets. Working off a budget acquired through sponsors and advertisers, SpaceFlight Insider has rapidly become one of the premier space news outlets currently in operation. SFI works almost exclusively with the assistance of volunteers.