The power inversion adaptive array

After the brief review on adaptive array processing, three fairly separate topics on the power inversion adaptive array are treated in this thesis. The first topic is the behaviour of a narrowband array using the stochastic gradient descent algorithm, with the environment assumed to rotate at constant velocity in the sine domain. Conditions for steady state weight deviations and output power deterioration from optimal values due to the nonstationary environment are derived and are then used to determine the maximum scan rate of a radar side-lobe canceller. The second topic is the jamming rejection capability of a broadband array using tapped delay line processing. The results obtained are used for designing the tap spacing and number of taps of the delay lines as well as assessing, in terms of the number of variable weights, the relative advantage of the alternative broadband processing method using several narrowband array processors. The frequency distortions at various directions introduced by rejecting the jammers are also studied qualitatively. The third topic is the convergence behaviour of the broadband array when the stochastic gradient descent algorithm is employed. Comparison with the alternative broadband processing method is again given. A simple transformation pre-processor, independent of the external environment and capable of improving the convergence behaviour of using tapped delay line processing, is also derived.