Three-dimensional antenna arrays are a less-explored family of aperture distributions when compared to linear and planar topologies. This is due in part to performance degradation from shadowing and scattering when the array spacing is fixed and on the order of a half wavelength, but the overall logistical complexity of their construction also creates challenges that have limited their practicality. However, when the spacing increases and becomes dynamic (e.g., a distributed swarm or cluster of UAVs) the array becomes sparse and unstructured. This may also include an unpredictable and morphing spatial distribution and orientation of elements in the array. All of this combines to create a complex three-dimensional manifold with unique performance attributes that enable the use of beamforming techniques not typically leveraged in uniform planar and linear arrays.
This talk will present the design and operation of Medusa 2.0, a new array test-bed designed to evaluate the performance of beamforming algorithms for stochastic array topologies with time-varying spatial distributions. It features a computer vision system to determine the position and orientation of (up to 32) elements that are used for hybrid (analog and/or digital) beamforming and algorithms to maximize SINR, SNR, etc. in the presence of interference. This includes a various GUI elements for visualization of the array and its directivity as the array morphs and/or the beamforming algorithm converges to an optimal solution. All of these steps take place in GNURadio, highlighting the ease of creating out-of-tree modules for many different applications that may not traditionally be associated with GNURadio.
|Talk Length||30 Minutes|
|Link to Open Source Code||https://github.com/HRG-Lab/gr-opencv; https://github.com/HRG-Lab/gr-medusa|