Edge ultrasound systems must sustain ultrafast imaging under tight power, memory, and I/O limits. This paper presents a plane-wave compounding pipeline that jointly compresses acquisition and beamforming: temporal undersampling exploits sparsity in radio-frequency (RF) data, while spatial compression replaces full-aperture delay-and-sum (DAS) with convolutional beamforming (COBA) and sparse variants (SCOBA/SCOBAR) that enlarge the effective sum-coarray using fewer active elements. The combined design provides axial/lateral resolution, contrast, and sidelobe suppression comparable to a full-rate, full-aperture baseline, thus enabling low-power, high-frame-rate imaging on portable devices. On experimental plane-wave datasets acquired from a calibration phantom at 4 MHz by 63-element and 127-element linear arrays (custom and commercial, respectively), the prototype sustains 50-60% temporal subsampling with 27-45 (custom) and 57-85 (commercial) active elements, improves contrast/signal-to-noise ratio (CNR/SNR) by 8.45/27.2 dB over DAS, and shrinks lateral resolution from 26.8 to 9.7 mm (FWHM). Computational overheads are modest, with SCOBAR-27 running within ∼1.3× of DAS runtime.
