This document summarizes an electrical engineering student's final project report on using an Unscented Kalman Filter (UKF) to estimate the state of a balancing robot. The UKF was able to accurately track states and uncertainties in simulation, but had difficulty estimating the true robot length from experimental datasets, likely due to insufficient oscillation of the robot. While the UKF and Extended KF agreed on the estimated length, more accurate methods may be needed such as Monte Carlo simulation with more samples.