obtain a satellite navigation receiver in the near future. Position infor- 

 mation can be entered into the computer either via the manual entry switches 

 at the computer site itself or from either of two remote input units. Upon 

 entry of the fix information the computer calculates the differences between 

 this fix and the dead reckoning position at the time of the fix, and records 

 this information on the main lab output typewriter and on the three remote 

 input- output typewriters on the bridge and in the science laboratories. 

 These differences are used in judging the quality of the fix in relation to 

 previous fixes, and in helping the operator to decide whether or not to use 

 the present fix for up-dating of the dead reckoning navigation carried on 

 by the computer program. Latitude and longitude, regardless of how it 

 is obtained, can be accepted by the computer. Computer programs have 

 been generated for the on-line processing of Loran-C and VLF micro- 

 second delays, and the latitude and longitude determined by these computer 

 programs are similarly recorded on the several typewriters. This 

 process is considerably faster and more accurate than the manual plotting 

 and reading of Loran positions. Similar programming for Loran-A micro- 

 second delays are in progress. The ship's position information processed 

 by the computer is also stored on a magnetic disk, and these data disks 

 maybe later reprocessed using off-line programs to up date the dead 

 reckoning navigation to agree with selected navigational fixes. Although 

 this procedure improves the continuity of position information along the 

 ship's track, it does not appear to significantly improve knowledge of 

 instantaneous ground speed. The ship's electromagnetic speed log is our 

 best available source of ground speed information. It will, of course, be 

 in error by any ocean currents that maybe present, sets due to wind, or 

 water turbulence along the hull of the ship. Surface ocean currents are 

 generally greater than 0. 1 knot and therefore our desired accuracy is not 

 obtained. However, there have been rare occasions when the computer 

 dead reckoning did not differ from navigation fixes by more than 3 nautical 

 miles for periods up to twenty hours. The determination of ground speed 

 indirectly from two navigational fixes may yield accurate values of 

 average ground speed if the time between the two fixes is sufficiently long. 

 Even if good fixes with an uncertainty of 0. 1 nautical mile are available, 

 at least two hours between such fixes are required in order that the error 

 in determining ground speed be within 0. 1 knot; and a two- hour average is 

 hardly an instantaneous value. Continuous 0. 1 knot accuracy in ground 

 speed does not appear to be generally possible at present. Even inertial 

 systems do not appear to be adequate for this accuracy, especially in 

 areas of rough bottom topography producing local deflections of the ver- 

 tical which cause erroneous accelerometer readings. We are presently 

 trying to improve on the existing situation by developing, in collaboration 



373 



