Maintained by: tpaine@mit.edu Get PDF
Learning Introspective Control (LINC)
This project aims to develop and demonstrate state-of-the-art approaches to control for underactuated marine vehicles which have experienced failures that result in significant changes to system dynamics. This project is a team effort led by the prime contractor, Aurora Flight Sciences (Boeing), and the Aerospace Controls and Active-Adaptive Control labs at MIT. We investigate the limits of classic control to handle significant disturbances in our Heron marine surface vehicles, and investigate the use of adaptive and learning-based controllers to provide better performance in such demanding situations.
In the second year of this project, the focus has shifted to investigating methods to assist human operators of marine vehicles with two challenging scenarios: Safely landing the vessel on a dock, and safely maintaining close position to another vessel for the purposes of transferring cargo. The primary test platform in the second year is our OPT-WAMV, a bigger vessel than the Clearpath Heron making it more capable in the open water.
Figure 1.2: Picture of OPT-WAMV used as the test platform in the second year of the DARPA LINC project.
In these tests the human driver of the vehicle provides a reference trajectory when close to the dock or other vessel, and the goal of the adaptive controller is to drive the vessel along that specified trajectory. The same adaptive control method that was tested on the smaller Herons is used. Collisions are avoided through the use of control filtering based on a constraint barrier function analysis. Virtual barriers are placed near the dock and other vessels, and if the operator's control input would result in the vehicle penetrating the barrier, the trajectory is adjusted to prevent unsafe situations. A simple diagram of this approach is shown below:
Status: | Ongoing since February 2023 |
Sponsor(s): | DARPA I20, Dr. John-Francis Mergen. Aurora Flight Sciences. |
People: | Tyler Paine, Mike DeFilippo, Mike Benjamin (PI) |
Robots: | https://oceanai.mit.edu/pavlab/wamv |
Software: | MOOS-IvP public codebase, MOOS-IvP-Pavlab codebase |
Recent Publications
2024 (2 items)
- Karan Mahesh, Tyler M. Paine, Max L. Greene, Nicholas Rober, Steven Lee, Sildomar T. Monteiro, Anuradha Annaswamy, Michael R. Benjamin, Jonathan P. How, Safe Autonomy for Uncrewed Surface Vehicles Using Adaptive Control and Reachability Analysis, 2024. (bibtex)
- Nicholas Rober, Karan Mahesh, Tyler M. Paine, Max L. Greene, Steven Lee, Sildomar T. Monteiro, Michael R, Benjamin, Jonathan P. How, Online Data-Driven Safety Certification for Systems Subject to Unknown Disturbances, 2024 IEEE International Conference on Robotics and Automation (ICRA), 9939-9945, 2024. (bibtex)
2023 (1 item)
- Tyler Paine, Michael Benjamin, An Ensemble of Online Estimation Methods for One Degree-of-freedom models of Unmanned Surface Vehicles; Applied Theory and Preliminary Field Results with Eight Vehicles, International Conference on Intelligent Robots and Systems (IROS), October, 2023. (bibtex)
Document Maintained by: tpaine@mit.edu
Page built from LaTeX source using texwiki, developed at MIT. Errata to issues@moos-ivp.org. Get PDF