Linux on a Manned Mars Rover

wstrong writes: "Lineo, makers of the realtime Linux uCsimm, has a story about a group of students building a manned Mars Rover. Based at The University of Michigan, they're using realtime Linux microcontrollers and "realtime" Ethernet to control the vehicle, from lighting and life support to drive-by-wire! Their first prototype, which is intended for operational testing on Earth, will be on display at Stanford in August for the Mars Society conference. They have raised almost $400,000 in donations for the project and its a high profile opportunity for Linux to break into space systems. Check the team's website for more."

Comments

Only 1 CS?

[heliocentric]

I looked through the list of students working on the project and I saw (pretty much as expected) a ton of EE and CE, but only one full CS (there is a CSE) major. It's usually been my experience when you get a gaggle of HW Engineers they tend to design really spectacular devices, but they tend to over look their CS counter parts and often work themselves in to a O(n^n) corner. I've also seen a "hey, let's do this with as much hardware as possible for logic" mentality and in many cases this is great, but it CAN be over done. IMHO, this looks like a great learning experience for all involved, and I feel this thing is Earth bound so my worries are probably moot, but whatever devices we send to Mars should be the best devices possible built by the best and designed by the best taking everything into consideration.

Re:Only 1 CS?

[freuddot]

This is a very stereotyped comment. I am a CE, and I've been programming console games for 3 years. And you wouldn't find any performance bottleneck caused by bad algorithms in my code. Most CE will have a natural tendency to concentrate on optimising specific code/instruction, I'll agree to that much. However, it only takes a run through the profiler to tells you that your bulleSort() is taking all the time. On the other side, most SE will have a tendency to implement this fast_hash_radix_sort_with_lazy_evaluation() only to find out three months later that a simple qsort() would have done the job perfectly. The important thing is to check what you *actually* need, and then get the people with the right knowledge (not diploma) to do it. And again, one profiler run is worth more than two SE and two CE thinking forever... ;-)

Good work and congratulations...

[Rick the Red]

...to the team from Michigan. But it's just another manned Earth rover. It won't be a manned Mars rover until someone sends it and a driver to Mars.