From: email@example.com (David A. Domino) Organization: MITRE Corp. Date: 30 Mar 96 16:01:10 References: 1 2 3 4
View raw article or MIME structure
In article <airliners.1996.378@ohare.Chicago.COM>, firstname.lastname@example.org (Joseph N. Hall) wrote: ----Snip---- > Also, there are the storage requirements ... let's say CONUS is roughly > 1500 x 3000 mi, sampled at 100ft intervals to a 16-bit accuracy ... > this means we need 4.5M x 52^2 x 16 bits = 22GB ... and how will all > that data be collected with verifiable accuracy? Why not just turn on > the radar altimeter and watch out for blinking red lights. Because the radar altimeter only reads what's below you and not what's ahead of you. In many terrain situations the rate of change of elevation, even if warning systems are working up to standard would be insufficient to prevent ground contact. A geographic database of _appropriately_ fine granularity would permit the design of a predictive system which could warn you in advance of ground contact. But your main point is that a comprehensive database might be prohibitively expensive, and in my view not particularly necessary to accomplish the task. For example there are pretty vast areas of the country where man-made obstacles would be the predominant risk, as opposed to terrain. In these areas the terrain data could be sampled at a much larger scale, and the obstructions database included with it. This would result in a much more manageable storage requirement, and still accoplish 99% (my WAG only) of the predictive terrain/obstacle avoidance task. Lastly, all of this technology is saving us from a fundamental communication failure, either between crewmembers in the airplane, between them and their automation, or between ATC and the airplane. When CFIT events happen, somebody didn't know something they should have, or could have known. A "relaxed" approach to the continuing question of "where am I, now" can't safely be tolerated, but the reliability of our modern systems may actually invite it.