UROP Openings

Have a UROP opening you would like to submit?

Please fill out the form.

Submit your UROP opening

Adaptive Controls for Autopilots for Rapid Aircraft Prototyping


Term:

IAP and Spring

Department:

12: Earth, Atmospheric and Planetary Sciences

Faculty Supervisor:

Sai Ravela

Faculty email:

ravela@mit.edu

Apply by:

02/01/2020

Contact:

Sai Ravela: ravela@mit.edu

Project Description

In this project, you will learn how to transform a stock Autopilot (PX4) into an autopilot that is resilient to system configuration changes, actuator and sensor anomalies by incorporating the retrospective cost adaptive control framework. The development part of this project consists of developing and testing dynamics and adaptive control algorithms in simulation (first) and in flight (second); we were asked to break things and push the envelope by our sponsor :) The research part consists of advancing ensemble approaches to model-predictive adaptive control, a new framework that using GPUs could allow for more complex and high-fidelity models to function in realtime for control. This project is part of funded research where you will join a group of people putting together new aircraft, new engines, and new dynamics and control algorithms for autonomous observation in hazardous environments. The Earth Signals and Systems Group has a rich history of supporting undergraduate research in vision, learning, autonomy and inference for the Earth and Planets.

Pre-requisites

You should know how to program in C, have experience with GitHub, and you should have taken at least one course in Controls. You should be interested in both dynamics and control of aircraft. Familiarity with the PX4 software environment and ROS will be a huge plus. This project will be perfect for a senior thesis and an MEng/SM thereafter but juniors and exceptionally well-prepared sophomores will also be considered. We request a commitment of 10 hours/week or better.