This project demonstrates the use of a particle filter for determining when a semi-autonomous vehicle should transfer control to the driver. I used two different particle filter distributions for each particle: normal and gamma. The particle filters were tested on a simulated semi-autonomous vehicle scenario created by Simian (Applied Intuition) to determine the optimal handover time. The reason particle filter was used instead of another belief-state algorithm is because the readings of any autonomous vehicle inevitably contain some noise. A gamma distribution was utilized because it is appropriate for model non-negative and skewed data, which is exactly the shape of the data measured by the sensors of the autonomous vehicle.