Can we predict the sea state using only on-board WEC motion and PTO measurements so that the controller can adapt in real time?

We train a neural network to infer wave parameters and excitation force online, enabling PI/MPC retuning without any external sensors.

MATLAB–Python co-simulation shows that this adaptive strategy can boost absorbed power by ~10–30% compared to fixed-gain control under changing sea conditions.

Publications: Energies (2025), IFAC-CAMS (2024), UMERC (2025)

What if every bump charged the battery? Can we harvest meaningful electrical energy from the suspension motion of electric semi-trucks as they drive over real roads?

We model the full truck dynamics and PTO in an electro-mechanical co-design framework, simulate diverse road profiles to quantify harvesting potential, and optimize the system to maximize energy capture while maintaining ride comfort and vehicle safety.

Publications: IEEE TEC (Under Review)

We developed an AI-based prognostics and health management framework for BLDC motors that continuously monitors vibration, temperature, current, and voltage under accelerated life testing to detect degradation early. Major contributions are:

1. Developed an NI-LabVIEW test rig for accelerated life testing, logging vibration, current, voltage, temperature, and speed.
2. Introduced a 3rd-harmonic current feature-selection technique, improving fault-diagnosis accuracy by 10 % over conventional methods.
3. Developed a remaining useful life estimation framework using machine learning to enable post-fault decision making.

Publications: IEEE Access (2020), IEEE Access (2021), IEEE Sensors Journal (2022)