Optimizing for Spatial Frequency Coverage vs. Point-Spread Function Sidelobe Level in Active Incoherent Microwave Imaging Arrays

Array optimization to maximize image reconstruction performance is often approached using numerical methods due to the solution space being too large for traditional brute force methods. In this work, a sixteen platform coherent distributed array with a seven element subarray attached to each platform will be optimized in two separate domains, each using a genetic algorithm: one to optimize spatial frequency content and the other to optimize peak to sidelobe level of the point spread function. The dimension of the array solution space is restricted to a planar domain of 50λ × 50λ, and the image reconstruction performance is compared for the outputs of both optimization approaches. It is demonstrated that optimizing the point spread function results in better reconstruction of images containing typical spatial frequency distributions, while optimizing the spatial frequency coverage results in better reconstruction of images containing mostly high-spatial frequency content, such as images consisting mostly of shape outlines.