M.Eng., Harbin Engineering University · Underwater Acoustic Engineering
Before Advanced Air Mobility, my graduate research at Harbin Engineering University was underwater acoustics: characterizing the transducers and hydrophones that generate and receive acoustic signals, reconstructing a real OFDM/QPSK communication link from an unlabeled lake-trial recording, building the detection-theory and array-processing foundations that both depend on, and using BELLHOP ray-tracing simulation to connect a canonical deep-ocean sound channel to real, measured shallow-water channel behavior. Four technical reports, each independently verified against real measured data, cited literature, and original from-scratch algorithm implementations, not reproductions of any textbook or paper.
Tank characterization of a piezoelectric transducer and hydrophone: electrical impedance in air vs. water, free-field voltage sensitivity calibration, transmitting voltage response, and a measured directivity pattern, each cross-validated against a computational model.
Engineered the exact modulation scheme of a real, undocumented 2018 underwater OFDM/QPSK recording from first principles, then used that model to measure real channel signal-to-noise ratio, coherence time, and a repeating synchronization pulse in the field data.
Connects DSP processor architecture, detection theory, and array signal processing to the measured results in the other two reports, with two original from-scratch demonstrations: a matched-filter detector and a MUSIC direction-of-arrival estimator.
BELLHOP ray-tracing simulation connecting the canonical deep-ocean Munk sound channel to a communication-relevant frequency case and a custom shallow-water channel, validated against independent physical and geometric expectations, then used for matched-filter signal detection over the resulting channel impulse response.
Every number below is measured or independently reproduced, not simulated for illustration alone. See each report for full methodology.
Measured beam width of 42.99° with an effective transducer diameter of 52.2 mm, matched against a computational model.
Real signal-to-noise ratio measured falling from 7.9 dB to 3.3 dB across the recording's 3-15 kHz band, consistent with underwater acoustic absorption.
Matched filtering later found a real synchronization pulse at correlation scores above 0.99 in the lake trial's own recording.
BELLHOP-computed transmission loss reproduces the Munk deep-ocean channel's real loss reduction, with every eigenray at 48 kHz terminating within 2 km and 50 m of the target receiver.
Impedance, calibration, TVR, and directivity analysis with a 3D computational model.
Versioned, citable snapshot of the codebase on Zenodo.
OFDM/QPSK scheme reconstruction and real channel characterization from field data.
Versioned, citable snapshot of the codebase on Zenodo.
Matched-filter and MUSIC demonstrations, DSP architecture background.
Versioned, citable snapshot of the codebase on Zenodo.
BELLHOP output-format parsers, deep and shallow channel scenarios, and matched-filter detection over the resulting channel.
Versioned, citable snapshot of the codebase on Zenodo.
Code mirror of the matched-filter and MUSIC demonstration scripts.