Slashdot Mirror
Spirit Rover Communications Error
cybrthng writes "Through yesterdays press release and the current Nasa Briefing there is news that they are having communications errors with contacting spirit. Is she lost or is it something akin to the Pathfinder failures that happened? Or did little green people claim an expensive tonka truck toy?"
But Spirit was only transmitting "pseudo-noise", a random series of zeroes and ones in binary code and not anything the scientists could decipher. - BBC News
It sounds like we still have power and an antenna. Hopefully its just some software error will need a reboot to correct the problem. I think they were late debugging this stuff and actually had to upload the software after the launch. Maybe they missed something.
The only issue I heard was some voltage spikes when the high-gain antenna was rotated. They were not reproduced but perhaps some underlying problem has occurred.
Up to now, NASA has made this look so easy. This is a wake-up call. Putting robots unto another planet is still an epic achievement and so much is left to go wrong even after the landing is over.
Let's hope this is just a red screen of death and a reboot will shake things loose.
By now they have probably rebooted it (forced it through safe mode to clear any software fault; space vehicles never really go all the way "down"), so if it's still happenning I would say it's either a hardware fault or corruption of essential software or data in (putatively) nonvolatile memory (not unreasonable in high-rad environments).
If it is corruption of secondary memory, and since they can send valid commands, presumably they can attempt to upload new data/code to fix it.
If they haven't forced it through safe mode, then they're not too worried and are more interested in characterizing the problem than getting on with the scientific mission. Which is a good or a bad thing depending on which sort of information is more valuable. I'm sure the guys in the software group have their bias.
Is it at all possible that getting half commands or garbled commands has confused Sprit?
They were just saying there are many sequences of events that could cause this. If it sensed the battery was overly discharged it will stop sending data & wait for a recharge. It could be as simple as this.
Two roads diverged in a wood, and I - I took the one the bus load of girls just went down.
Or, more clearly, what do the operating routines look like? Does anyone have a flowchart that would show the data flow? What sort of error checking is done on incoming data? What sort of encryption is done on incoming/outgoing data? (Cartoon bubble: I picture a script kiddie with a powerful transmitter sending SQL injection to Spirit...)
What does the system do if it determines it has had an unexpected result/crash? How is such a system designed and tested?
I've never thought about it before - but a system like this must have redundant levels of i/o security, internal error checking, exception trapping, and some sort of self-repair, all built within multiply redundant systems.
Would any details of the embedded system architecture / program structure be available to the public?
The cure for cancer is coming: Reovirus
- Part 1
- Part 2 (more detail)
There's a lot of rumor and inconclusive news about Spirit floating around right now, so this is entirely subjective, but I'm getting the feeling this, too, is a software fault of some kind. Put most simply, you could interpret what we're reading right now as "we received the ACK tone for our instructions but didn't get the data back we expected."These kinds of problems are not unprecedented, and furthermore I'm under the impression there are options for dealing with even serious OS-level trouble that would shock and awe the average general purpose computer user.
Want to Know How to Cheat the GPL? Read On!
...to have an engineer on scene to fix it.
Robots are great until they break.
-- Slashdot: When Public Access TV Says "No"