Enhanced Range Alignment Using a Combination of a Polynomial and Gaussian Basis Functions

Abstract

Abstract-For the inverse synthetic aperture radar (ISAR) imaging of a target at a long range, range alignment using the existing polynomial method brings about poor results because the flight trajectory changes depending on the initial position, and the motion parameters, meaning the polynomial cannot fit the trajectory. This paper proposes an improved range alignment method that models the trajectory using a combination of a polynomial and Gaussian basis functions. Initial parameters of the polynomial and Gaussian basis functions are determined by fitting the proposed model to the center of mass curve of the range profile history using the least square curve-fitting algorithm, and the optimum value is found using particle swarm optimization. This method is computationally more efficient and preserves the image quality.

An Efficient Isar Imaging Method for Multiple Targets

“…This method polynomial was applied to range alignment of a single target using the maximum energy [14] and was less accurate than the nonparametric method [15]. However, a combination of this polynomial and the cost function defined in [10] effectively aligns RPs of multiple targets and generates the bulk image.…”

“…In this paper, we model the target trajectory by using a combination of a polynomial and Gaussian basis functions and applying the estimation procedure as given in [14]. Given M range profiles each having N range bins, the function that represents the target trajectory is expressed as…”

“…However, a combination of this polynomial and the cost function defined in [10] effectively aligns RPs of multiple targets and generates the bulk image. For the estimation of the parameters in (1), we apply the gradient-based method in [14]. In setting the initial values of the parameters in (1), we use the parameters that fit (1) to the COM curve of the range profile history.…”

Efficient Range Alignment Algorithm for Real-Time Range-Doppler Algorithm

“…In real imaging situations, a long flight trajectory is required to obtain the aspect angle needed for a given Doppler resolution; as a result, the number of range bins increases drastically, and the trajectory can be highly nonlinear. Consequently, the computation time of the methods in [4] increases, and the alignment results of [5][6][7] based on the trajectory modeling degrade due to mismatch between the trajectory and the polynomial function that is used to approximate it. The algorithm in [8] also consumes more computation time because of the poor convergence of the cost function.…”

“…Some derived methods [5][6][7][8] reduce the computation time and increase the accuracy of alignment. However, the main problem of existing RA methods is that the results were mostly obtained when the flight trajectory of the targets was very short and nearly linear.…”