Adaptive nulling in monopulse antennas

Theoretical and experimental results are presented for phase-only nulling in low-sidelobe monopulse antennas. Both results are based on a gradient search algorithm that simultaneously searches for a minimum in the sum and difference channel output powers. The array's beam steering phase shifters double as the adaptive weights. Each element in the gradient is found by changing phase shifter setting by Delta Psi (the phase shifter stepsize) and measuring the change in output power. Then the phase shifter is restored to its original value, and the process repeated for all the remaining array phase shifters. The algorithm iterates as long as each new adaptive weight setting reduces the total output power. If the output does not go down, then Delta Psi is decremented by one setting and the iteration is started again. The algorithm stops when Delta Psi =0. The adaptive weights act as random perturbations to the phase taper of the array. Consequently, the sidelobe level is proportional to the size of the phase perturbations and inversely related to the number of elements. By keeping the adaptive phase shifts small, the average sidelobe level and the main beam gain do not drastically change. >