Using a leaky-wave antenna as an acoustic prism for speech-separation

Abstract

Leaky-wave antennas (LWA) are an acoustic metamaterial comprised of an array of unit cells. For a 1D line array, the output ranges from 0° to 180° relative to the direction of the axis of the array. The direction of the output is dependent on the frequency of the input; a single frequency input will produce a highly directive output beam at a specific angle. Much like an optic prism, which can be used to split a complex optic wave into its components, a LWA can be used to split a complex input audio wave into its components. This work explores the capabilities of LWA to separate transient signals like human speech into different output directions. This is accomplished by way of an equivalent circuit model of the LWA that enables simulations in both the time and frequency domains. The frequency domain performance of the model is validated against previously published results, and then the transient response is shown to recover these as well once steady-state is reached