Re: What language was used on Deep Space 1?

[Ted Dennison <dennison@telepath.com>]

In article <37A12E5C.DD1F8720@easystreet.com>,
  Al Christians <achrist@easystreet.com> wrote:
> JPL's Deep Space 1 probe just did quite well using an advanced
> autopilot.  JPL did use Ada on their Mars mission a little while
> back.  Did Deep Space 1 use it too?  Can anyone here tell how this
> software was built?

It just so happens that I have a paper here that was required reading
for my AI class titled "Smart Executives for Autonomous Spacecraft",
written by Erann Gat of JPL and Barney Pell of NASA Ames Research
Center. It was published in the September/October 1998 IEEE Intelligent
Systems, for those who may be interested.

The paper discusses the development of an executive for an autonomous
spacecraft, dubbed "Remote Agent". According to the article, "RA's
implementation language is Common Lisp. RA runs in Allegro Common Lisp
under Unix and on the flight processor (a Rad6000--radiation-hardened
processor similar to a PowerPC) under a custom port of Harlequin Common
Lisp for the vxWorks real-time operating system." It appears that some
sub-components of the Remote Agent were written in their own modelling
languages that were built upon lisp.

Since this is primarily an expert system problem, its not at all
shocking to find that it was written in Lisp. I've gotten farily
familiar with how forward-chaining inference engines operate thanks the
the class project I've spent the better part of this week working on. It
would be a fairly difficult thing to program directly in a procedural
language.

However there are some similarities between how an inference engine
works and how a compiler works. Both are basicly in the business of
matching expressions. I've given some thought to one day seeing if the
OpenToken approch to providing compilation services to an Ada program
could be used in the expert system domain as well.

Anyway, it appears the project's homepage is
http://ic-www.arc.nasa.gov/ic/projects/Executive/ . There are links to
lots more papers there if you are interested.

--
T.E.D.


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.