Calibration and fusion of accelerometer, gyroscope and compass data in mobile and gaming devices.

My PhD thesis focused on digital adaptive control algorithms used for both control and estimation. In particular, I ran millions of Matlab simulations to quantify the stability of an in-house developed adaptive control algorithm. I then improved this version of the algorithm by imposing a convex constraint on the controller poles, which decreased the transients by 70% and settling time by 50%. Lastly, I explored the nonlinear extensions of this algorithm by applying it to nonlinear and time varying systems like multi-link robotic arm.Another aspect of my graduate work was in modification of the ensemble Kalman filter to estimate states and drivers of an ionospheric model. My main modification included introducing ability to estimate model inputs to the system in addition to model states. The ability to estimate unknown drivers is extremely useful in practice if, for example, one desires to get estimates of acceleration from noisy accelerometer measurements while using Newton or Euler equations with unknown force and torque inputs. In addition to utilizing the estimation theory, this project required a lot of software engineering, as I wrote the Fortran interface between the model and the ensemble Kalman filter and analyzed the results of assimilating GPS satellite data using Matlab and Python's Matplotlib.

While being a part of a student team building a fuel-efficient vehicle for the SAE Supermileage design competition, I built and programmed an Atmel AVR microcontroller based GPS-accelerometer-gyroscope system to control the engine. After selecting the hardware and coding up the control law in assembly and C, I installed the embedded system on the vehicle, which along with other modifications improved the fuel efficiency by 250% (from 456 MPG to 1140 MPG) and helped our team advance from 9th place the year before, to 2nd place.

Participated in project design management, project reporting, and led a 35 student member team during first participation efforts, achieved 27th place out of 64 in a national floating all-terrain vehicle competition. Upgraded previous project model design frame to improve functionality, length and weight for competition and safety requirements. Drove project vehicle for actual race during completion.