Kaplan, Sargent and Goodman 
the ratios C,/C, and C, /C, , which is the form in which these 
coefficients are determined, are about the same from the estimates 
and the true values. In general the values shown in Table 3 are suf- 
ficiently close to represent a successful identification of the system 
parameters, as is also indicated by the trajectory matching in 
Figure 9. 
Full scale responses of a hydrofoil craft (essentially the 
same craft treated above using simulated responses) were provided 
as a means to test the techniques developed herein for system iden- 
tification. This data was supposed to be obtained as a result of im- 
posing various commands to the craft controls (i.e. elevator and 
flaps) in an ostensibly smooth seaway, i.e. they were expected to 
represent transient responses of the craft. The basic equations re- 
presenting the craft motions included the additional degree of freedom 
due to surge; the craft propulsive thrust force was represented in the 
equations as a function of propeller rpm, and drag forces (which 
included quadratic terms in resulting angle of attack) were also pre- 
sent in their influence on surge and also the pitch moment (see (1 0] 
for a more detailed discussion of these equations), The initial intent 
was to apply the iteration method for transient responses that was 
discussed previously, and this approach was attempted. The equations 
were simplified to eliminate the surge degree of freedom (since the 
speed was almost constant during these particular maneuvers) and the 
iteration method was applied. The results showed a lack of conver- 
gence and/or stabilization throughout many attempts, and further 
simplification was made to eliminate the drag forces from conside- 
ration since the motions were generally small. 
For the more simplified representation of the craft 
dynamics success was still not achieved due to lack of convergence 
and/or stabilization of the parameter estimates, although for some 
portion of time the estimated trajectories seemed to be in fair coin- 
cidence with the measured data. The experimental trajectory data, 
which included the pitch angle, pitch rate, and the height sensor time 
histories, did exhibit fairly large ''noise'' superposed on the main 
craft responses due to the commanded control deflections. A typical 
illustration of the control deflections experienced in the full scale 
tests is shown in Figure 10, and similar types of noise were present 
in all of the recorded data (the observed data from the tests was 
presented in digital form every .05 sec., asa result of digitizing 
the measured craft motion tape records). The record for the flap 
deflection really represents the average of starboard and port aileron 
deflections, which was the chosen method of representing this control 
action in the analysis of this trail data. Simple smoothing operations 
1662 
