Dominance of stiffening effects for rotating flexible beams

beam cycle, therefore A varies by 72 deg between data. The initial estimates (f,0) converge toward the actual values. The noise in the data is simulated as having a constant covariance, which may not be the case . If a dependence of noise on range can be estimated or conjectured, the effects of such noise on the estimate can be forecast. Figures 2 and 3 show the time variations of errors in the estimates of range and azimuth angle, with their rms errors (la). Despite large nonlinearities and an indifferent initial estimate, the convergence is rapid arid the errors (dr,d0) are consistent with the estimated standard deviations. In an actual application, of course, the errors are not known and the standard deviations are found, as in this example.