Synthesis of Conical Phased Antenna Arrays Optimization of Amplitude Distribution Parameters

Abstract

In the synthesis of conical phased antenna arrays (CPAA), special attention is paid to its characteristics a high antenna gain and a low side lobes level (SLL). To solve this problem there are a significant number of approaches, in view of the variety of methods of its formulation and the methods of solution. In the present paper, the problem of synthesis of a conical phased array is solved by optimizing the parameters of a two-parameter family of amplitude distributions by the criterion of maximum gain with the limit of the maximum side lobe limited by the methods of random search (based on a swarm of particles) and classical gradient methods. To calculate the loss function, a fast Fourier transform is used. Introduction The phased antenna arrays (PAA) with the radiating surface in the form of circular cone allow making undistorted circular scanning of the directional pattern in azimuth and wide-angle scanning in place angle that are of interest to the systems of radar-location and communication. One of the most important problems in the pro jection is the assurance of antijamming and electromagnetic compatibility with other radioelectronic resources. The desired result can be achieved by synthesizing a CPAA with a limited side lobes level and a high gain. In the synthesis by the method of numerical optimization of a phased array, the speed of the algorithm for calculating the directional pattern (DP) plays an important role. In this work, to reduce the synthesis time, an algorithm was of a fast Fourier transform (FFT)[2] was used. In the synthesis of CPAA, an amplitude distribution was chosen that depends on two real parameters p and γ, which allows adjusting of the side lobes level and the width of the main radiating beam. In this paper, one of the ways of setting and solving an optimization problem for obtaining the CPAA parameters with a high gain value under the constraint of SLL is investigated. The optimization problem was solved by random search methods and gradient methods, and the results were compared. The choice of optimization methods is explained by the successful application of random search algorithms[6] convex programming methods in a similar problem [7]. 1 2