Thursday, December 01, 2005
Galileo's Success is due to Redundancy & Flexibility in its Design
Photo Courtesy NASA/JPL-Caltech
Wow! I'm impressed with the Galileo Group's research into the Galileo mission.
Here a picture of the surface of Europa (second closest major moon of Jupiter). Apart from other evidence, this image shows why scientist believe there's an ocean beneath Europa's surface. What do you think it shows?
This may be beyond the scope of the Lego Ocean Odyssey challenge, but its something you may want to think about when you are designing your robots.
The success of the Galileo mission at Jupiter was due to 2 important design elements that everyone should try to plan into their robot designs. Galileo had redundant systems for critical functions and a versatile computer that was capable of being reprogrammed from Earth.
The Galileo mission was designed to send high volumes of data over its high gain antenna from Jupiter. This antenna was suppose to unfold like an umbrella, from its stowed position after the spacecraft was on its way to Jupiter. However, a few of the pins that held the ribs of the umbrella-shape antenna became permanently stuck and thus the antenna could not open. Fortunately, Galileo also had a redundant low gain antenna, but the rate of data sent back to Earth was only a trickle. The comparison is like the slow drip of a faucet to buckets full of data being sent back to Earth.
To make use of this 'slower' antenna, engineers had to redesign the entire mission with this lower data rate in mind. They programmed algorithms to compress or shrink the size of the images without losing too much information. This helped Galileo send thousands of pictures which otherwise would have been impossible. Furthermore, with the high gain antenna, the spacecraft would essentially relay the images to the ground as they were being shuttered. However, this would not work with the low gain antenna. For the low gain mission, engineers made good use of Galileo's tape recorder. Data from the close observations of the Jupiter's moons were compressed and recorded on the tape recorder. Then the data was read out into the telemetry stream and modulated onto the radio signal sent back to Earth. So the tape recorder was essentially a redundant system for handling the data and its return to Earth.
Now, with the tape recorder becoming a vital part of the redesigned mission, we could not let it fail. Our precise navigation requirements for entering into orbit around Jupiter required us to take optical navigation images of Jupiter's moons. But on the final approach to Jupiter, the tape recorder's tape became stuck. I believe it was pretty much like glued to the tape head. Without these images, Galileo's orbit insertion did not meet the requirements needed to stay on our planned orbital tour of Jupiter. Faced with this possibility beforehand however, we came up with a plan to change the first orbit around Jupiter according to the miss distance from our target and eventually reconnect the orbit to the planned tour. A few months after orbit insertion, the engineers were able to figure a way to unstick the tape and get it working again. They had to be very careful since it would have been very easy to break the tape. Luckily, they had a spare tape recorder at JPL to verify the problem, test their ideas on and verify the fix. According to my memory, this problem happened 3 or 4 more times during Galileo's mission.
I won't go on and on, but there were a few other major problems that Galileo experienced. Despite these problems, engineers and scientists were able to overcome them by making use of Galileo's redundant systems and by reprogramming it from many 100's of millions of miles away. When you can only afford to build one spacecraft and send it to Jupiter, you need to provide your spacecraft with backup systems and allow flexibility in the design of its computer and how it interacts with all its subsystems to control the vehicle.