Kaplan, Sargent and Goodman 
where u = surge velocity, 
v = sway velocity, 
r= yaw rate, 
6 = rudder deflection, 
f(u) = known function of u that represents the difference 
between thrust and resistance, 
C.= coefficients to be determined. 
These equations were obtained from (11), and the ''measured" input 
to the system identification program is provided by a maneuver 
generated on the computer with known parameters typical of a con- 
ventional cargo ship, as obtained from [1 1 . The maneuver chosen 
is a turning circle initiated by commanding a 35 degree rudder angle. 
Difficulty in converging to the known set of eleven coefficients 
was encountered with the basic computer algorithm even though the 
predicted time histories of u, v and r themselves became indis- 
tinguishable from their respective input records, Efforts to improve 
the results by sampling at a higher rate or by taking a longer record 
proved fruitless, and only by retaining an inordinate number of 
significant digits in the input data did the converged coefficients agree 
with their known values. However, the accuracy of data measured by 
real sensors is limited, and, the difficulty in obtaining good coeffi- 
cients was eliminated by applying the modification to the technique 
which used the measured state variables instead of the predicted 
values in the operations of Equations (1) and (6). Using this modi- 
fication, the effect of varying the sampling period, length of record, 
number of records, and accuracy of the data was investigated to 
indicate the measuring requirements for the identification of a real 
ship (or model) from a maneuver. It was found that a sampling period 
of from one to two seconds and a record length of from one-half to 
one minute was adequate for successful identification of a full scale 
ship. However, it was found necessary to have measured all three 
variables, namely u, v and r, to identify the eleven coefficients in 
Equation (13). 
Using the computer generated data of u, v and r accurate 
to four significant digits, which was sampled every second for one 
minute, the identification by this iterative technique was carried out. 
The resulting C-coefficients are shown in Table | together with their 
time values, where reasonably close values were found. The identi- 
fication process converged very rapidly, as only two iterations were 
required. 
1640 
