Reduceing the size of MIMO arrays using adjoint-based geometry evolution

This paper presents an approach for reducing the size of multi-input multi-output (MIMO) system antennas. In this approach, we minimize the spacing between elements such that both the return loss of each element and the coupling between any two elements are less than -10 dB. To reduce the spacing, the shape of individual antenna elements is allowed to evolve using a gradient-based optimization. Adjoint sensitivities are utilized to efficiently solve the optimization problem. We illustrate this approach by designing a 1-by-2 ultra-wideband (UWB) microstrip MIMO array.