MERS (Model-based Embedded Robotics Systems) is a lab at MIT CSAIL that uses algorithmic approaches to more efficiently and flexibly plan and deploy robotic agents.
Our lab specializes in model-based heuristics to solve problems like:
Dispatching for long-duration missions
How do we ensure an autonomous agent can precisely and confidently achieve a mission under unbounded mission duration like exploring on the surface of a planet?
Planning under uncertain time constraints
How do we dynamically schedule our agents to accomplish thier missions despite changes in the environment and uncertainty on when events will occur?
Risk-aware agents, safety, and LLMs
How do we make agents act with a risk-aware mindset and leverage state-of-the-art language models to simplify the planning process?