Tensor Beamforming for Coprime Planar Arrays With Virtual Peak Suppression

The patent transforms conventional vector signal processing into tensor space processing by introducing a new dimensional framework. The receiving signal is modeled as a fourth-order tensor that preserves multi-dimensional spatial information (DOA and polarized state) across different dimensions, rather than flattening it into a simple vector. This dimensional elevation allows the processing system to maintain original structural information while working in an expanded mathematical space.

The patent employs composite tensor beamforming that combines multiple processing approaches: it integrates tensor modeling with conventional beamforming techniques, and further combines results from multiple coprime sparse uniform sub-planar arrays through composite processing. This composite approach leverages the strengths of different methods while mitigating their individual weaknesses, particularly in preserving structural information.

The patent divides the electromagnetic vector coprime planar array into multiple coprime sparse uniform sub-planar arrays. Each sub-array is processed independently to form its own tensor beam power pattern, and then the results are combined through composite processing. This segmentation allows the system to exploit the sparse arrangement benefits (larger array hole diameter, higher spatial resolution) while managing virtual peak interference through individual sub-array processing.

The patent converts the virtual peak interference problem into a beneficial feature by designing the coprime sparse uniform sub-planar arrays such that their virtual peaks occur at different positions. When the tensor beam power patterns are combined through composite processing, the virtual peaks of different sub-arrays do not coincide, effectively suppressing the harmful interference while maintaining the spatial resolution benefits of sparse arrangement.

The patent segments the complex high-dimension tensor beamforming weight design into multiple manageable components. Instead of designing one large adaptive weight tensor, the system designs separate weight tensors for each coprime sparse uniform sub-planar array, and then combines them through composite processing. This segmentation reduces the computational burden and design complexity while maintaining beamforming performance.

The patent handles the high-dimension tensor beamforming weight by working in tensor space rather than conventional vector space. The fourth-order tensor structure allows the system to organize and process the complicated space information (DOA and polarized state) in a structured manner across multiple dimensions, making the weight design more manageable despite the increased dimensionality.