Good knowledge of Python, some knowledge of machine learning
Drones are cool, but sometimes things go wrong. Imagine a drone is exploring an area, and all of a sudden the GPS signal becomes unreliable. How can the drone estimate its location, and avoid getting lost? Perhaps we can use data from its sensors and picture from the camera.
In this project, we would like to study how machine learning algorithms could be used to provide useful information to the drone regarding its position.
You will use the popular AirSim simulator (https://microsoft.github.io/AirSim/) to simulate a drone operating in a known environment. Here is a video of the simulator:
You will collect data from cameras and other sensors, generate data sets, and then feed them to machine learning algorithms to estimate the drone’s position.
Possible follow-ups of the project include a hardware implementation and field tests of the system, in collaboration with the drone unit of the cs department.