Compact bandpass pixelated microwave filters with short-circuited stubs via inverse design

Pixelated RF metasurfaces are poised to revolutionize electromagnetic component design by enabling compact, versatile, high-performance solutions. Building upon our prior work in random metasurface-based filters and inverse design methods, we propose pixelated notch filters by integrating shorted stubs within a top ground plane. Using a combination of established optimization techniques, including direct binary search optimization, genetic algorithms, and a randomization mutation algorithm, we synthesize filters enhanced by parallel short-ended feed schemes, which are shown to improve stopband response. Design iterations are automated via Python scripting, commercial full-wave simulations, and Visual Basic within the electromagnetic solver, overcoming initial seeding challenges and enabling innovative pattern-generation techniques. For implementation, laser ablation is employed to precisely remove copper on PCBs, streamlining fabrication on Rogers Kappa 438 substrates. Preliminary results demonstrate the ability of the approach to reach target insertion loss levels with compact geometries, advancing pixelated metasurface-based filter design with enhanced tunability and overall performance.