Mars Rover Upgraded

MrShaggy writes "According to a BBC article, NASA is upgrading their MARS rovers. The upgrade will allow the rovers to sift through the pictures of dust-devils, decide which is the most appropriate, send it back. 'Clouds typically occur in 8-20% of the data collected right now,' Castano said. 'If we could look for a much more extended time and select only those images with clouds then we could increase our understanding of how and when these phenomena form. Similarly with the dust devils.' The article also discusses upgrades to the Mars Odyssey. They plan to make it self-reacting to events on the planet as they are happening."

Old News

[maytagman]

I heard this reported on CBC radio SEVERAL months ago. I'm thinking it was febuary... The scientist they were interviewing was saying how hard it is to trust a robot to make the right decision even though they knew the algorithm they were using was pretty fool proof. Lets hear it for CBC radio!!!

Re:Rumors that they're 'upgrading' from Ada.

[Anonymous Coward]

Yes, they'd have to write their own JVM. They aren't the only ones who do this, www.pilz.com do the same for industrial software.

Re:Rumors that they're 'upgrading' from Ada.

[Avionics Guy]

The large majority of the MER software was written in C. The exception is a small module in the navigation code that used C++ with a custom memory manager. BTW, JPL doesn't "do" ADA and it isn't likely that Java will be used on the MSL, the 2009 rover.

Re:Rumors that they're 'upgrading' from Ada.

[MWales]

I believe the current ones would probably use C/C++ since they are using VxWorks according to Windriver [windriver.com]. If they are using a RTOS now, I think moving to something like Java would be a huge jump. I could see them moving to embedded Linux though, it's becoming alot more popular in the embedded world

Re:Absolutely amazing

[kilodelta]

If you want, read Steve Squyers book "Rovign Mars". It'll give you a better understanding of why the rovers lasted as long as they did. They're built like tanks with proven technology. There was nothing flashy about what went into those robots, it was all tried and true.

They were originally supposed to last for 90 sols, or Martian days. They've now gone far past the origianl design goals and the benefit has been lots more data about Mars. Spirit is currently on it's 853rd sol. http://marsrovers.nasa.gov/home/ [nasa.gov]

Re:What Upgrade?

[the_brobdingnagian]

The upgrade is a software upgrade. But it's not an easy task to do this at such a distance. Two way communication is a painbecause of the lag time. I can't remember the exact time, bu I believe the lag is about 20 minutes. They use a specialised protocal that was designed to handle such extreme lag. The protocol is PROXIMITY-1 SPACE LINK PROTOCOL (specs [nasa.gov]). They are verry carefull to make sure they dont have to reset the rover the hard way (A.K.A. reset-button) after updates and even during normal operation. I believe they build in all kinds of auto-reset features so the rover could reset itself.

Not a PR conspiracy

[Tablizer]

They were built with the idea that they could conceivably last this long but the mission profile (and all the press releases) were put together with the expectation that they'd last a couple months. It was the closest thing to a gaurenteed win NASA could do. Think of it this way, if GM marketed...

Hogwash. It is a combination of factors:

1. Nasa increased quality control effort and spending in response to the Polar Lander failure and two orbiter failures.

2. Wind has blown dust off of the solar panels. Many expected the dust to be probe-sticky and accumulate based on the Viking lander data.

3. Constructor contract payments were actually stipulated based on a 3-month survivle. It is not an arbitrary deadline.
       

Re:Rumors that they're 'upgrading' from Ada.

[glwtta]

Even considering the Arianne-5 failure, it's still more reassuring to know that a software system is developed in Ada than Java.

Arianne 5 was the result of pure, old-fashioned incompetence. An obsolete component - left on when even its original function would not have been needed - dumps debug info on the bus, that's then interpreted as trajectory data. And the backup system runs identical hardware and identical software to the primary (I believe the backup actually failed a fraction of a second before the primary).

The rover software on the other hand - written in C, btw - is a gold standard of excellent engineering and testing practices. Most of the time it's not the platform that counts, it's the development team.