Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site ucla-cs.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!vax135!cornell!uw-beaver!tektronix!hplabs!sdcrdcf!ucla-cs!lmiller
From: lmiller@ucla-cs.UUCP
Newsgroups: net.aviation,net.graphics
Subject: Re: Flight Simulators
Message-ID: <6758@ucla-cs.ARPA>
Date: Mon, 2-Sep-85 13:14:05 EDT
Article-I.D.: ucla-cs.6758
Posted: Mon Sep  2 13:14:05 1985
Date-Received: Thu, 5-Sep-85 07:42:36 EDT
References: <1112@mhuxt.UUCP>
Reply-To: lmiller@ucla-cs.UUCP (Dr. Lawrence Miller)
Organization: UCLA Computer Science Department
Lines: 42
Xref: watmath net.aviation:1958 net.graphics:1075

In article <1112@mhuxt.UUCP> evans@mhuxt.UUCP (crandall) writes:
>Have there been any developments in flight simulators...
>Any comments on what is minimal hardware--
>is an Amiga with its graphics hardware in the ballpark?

During the past 18 months we have been building/writing flight simulators
using IRIS terminals and workstations, primarily for investigating pilot-
aircraft interfaces and expert systems for aircraft and pilot performance.
We've added joysticks, throttles, voice I/O, and lots of tools for rapidly
interacting with the display tables.  The screen resolution is 767x1023, 12
bits per pixel, double buffered (leaving 6 per swappable image, less 2
reserved for certain background images, or 16 colors on the screen at any
one time).  This number of colors is marginally adequate.  The resolution
is fine.  We do a substantial amount of non-graphics calculations per
display loop, so we're updating the image between 5 and 10 times per
second--not quite adequate for smooth motion.  Ten updates per second is
the minimum you'd want.  Obviously, we do lots of tricks to get as much
speed as possible, but other issues are more important now.

We designed and built our own joystick/throttle and interfaces, using one
of the RS-232 ports on the IRIS.  Our joystick is a 9600 baud ASCII device
that sends its x, y and throttle values (using A to D's and an 8008) when
queried by the program.  The joystick itself was obtained from a used
aircraft parts shop, but there are new ones with all the A/D's built in
that are not very expensive.  Or you can go to McFadden controls and get
real simulator hydraulics for 50K each.  Plus an oil change every 1,000 hours.

You will tire of slow, low resolution, poorly modelled flight quickly.  An
IRIS or equivalent (raster tech, chromatics, etc.) with fast graphics
computing is the minimum you will want.  There is still lots of other
computations needed.  The IRIS workstations are in the 40-60 K region, as I
recall.  Aerodynamic simulations can be reasonable with only very simple
models.  It is not necessary to model all the flow equations, since they
turn out to be inaccurate anyhow, and it is the solutions you want.  For
special circumstances, such as ground effect, simple heuristics are best.

By the way, we are interested in hearing from experienced UNIX/C
programmers, with extensive graphics background, who also have a pilot's
license.  The work is being done at ISI.

L. Miller
{lmiller@isi-hobgoblin or lmiller@ucla-locus}