capable of cutting operations controlled by a tape or other input program. 
The programed instructions could either be inserted at the surface before the 
machine is lowered, transmitted sonically once it is on the bottom, or perhaps 
an optimum combination of the two methods could be used. From a cost 
standpoint, it appears simplest to preprogram the machine. However, this 
would permit no change in plans and would require precise positioning. 
Either a change in plans or imprecise positioning of the machine in the work 
site would make some sort of corrective action necessary. Several approaches 
appear possible: (1) use of a telemetry link and reprograming from either a 
submersible or from the surface on the basis of closed circuit TV information, 
(2) replacement of the program by a submersible, or (3) orientation correction. 
(The third alternative would not solve the problem of a plan change.) 
The remotely instructed machine performing planned excavation 
and drilling tasks should take advantage of the latest technology in the 
automatic machine tool industry. The controlled production machine is 
avery old concept; probably the first example was the Jacquard loom, con- 
trolled by the equivalent of punched cards. Progress has been from punched 
cards through digital tapes and more recently to direct computer control, 
using a formal machine language. This approach is seen to have the greatest 
potential for the application considered here, with the most sophisticated 
systems being able to read a simple blue print of the site plan and perform 
the various necessary operations in a planned routine. For instance, a com- 
plex site might consist of a series of foundation trenches, a level table area 
and a series of holes for pipe pilings. The machine might level the site first, 
then automatically go into a trenching routine followed by the drilling of the 
vertical holes. Fluidic control devices of the type investigated at the Naval 
Ship Engineering Center, Philadelphia (Wexler, 1969), should be particularly 
useful in developing the numerical controls seen to be desirable. 
In summary, excavation at the extreme depths to be encountered in 
the DOT program probably will not allow for a man-controlled machine. 
Specific excavation functions can probably best be controlled by the numer- 
ical and computer techniques used with precision manufacturing machines, 
especially the milling machine. Optimum use of hydraulics, including fluidic 
control devices from the first should reduce cost and increase reliability. The 
technology is available for adaptation. 
The results of a major analysis and preliminary design (Northrop, 
1970) indicated that there might be a major technological deficiency in the 
present ability to accurately position a cutter in a real-world domain. 
Approaches and possible future work in this area are discussed in ‘’Shallow 
Drilling and Excavating’ of this report. 
69 
