Projects
Selected research and engineering projects across surgical robotics, medical sensing, machine learning, and embedded rehabilitation systems.
- Trained SLEAP-based deep learning models to track feature points on da Vinci surgical instruments.
- Applied SolvePnP to estimate surgical tool poses from visual observations.
- Developed particle-filter fusion of computer vision and forward kinematics to compensate for hand-eye calibration errors.
- Improved pose estimation robustness during tool-environment interaction.
Computer VisionSLEAPSolvePnPParticle Filtering
- Modeled curved flexible magnets using magnetic dipole superposition.
- Implemented constrained Kalman filtering for magnet localization and orientation estimation.
- Reconstructed continuum robot shapes using interpolation and curve-fitting techniques.
Medical RoboticsMagnetic SensingKalman Filtering
- Developed Python pipelines to extract gait characteristics from wearable sensor data.
- Applied ReliefF and KNN-based feature selection methods.
- Built neural-network models in MATLAB for muscle fatigue prediction.
- Conducted human-subject experiments and data collection.
PythonMATLABNeural NetworksWearable Sensors
- Built an OpenWRT-based communication gateway on Raspberry Pi 4B for data transmission between devices and host computer.
- Developed an interactive embedded interface using ESP8266 and touch screen.
- Designed and fabricated shells for EMG and inertial sensors.
OpenWRTRaspberry PiESP8266Embedded Systems