Matt Foutter

Papers

Total Citations

H-Index

About

Matt Foutter is a researcher at the forefront of integrating large language models (LLMs) into robotics, specializing in real-time anomaly detection and reactive planning. His most-cited work, "Real-Time Anomaly Detection and Reactive Planning with Large Language Models" (2024, 28 citations), pioneers the use of foundation models to identify and mitigate out-of-distribution failures in robotic systems—a critical challenge for deploying autonomous agents in unpredictable environments. By leveraging LLMs’ zero-shot generalization capabilities, Foutter demonstrates how internet-scale pre-trained models can detect anomalies without task-specific training, enabling robots to adaptively replan in real-time. This work bridges the gap between static robotic control and dynamic, real-world safety, offering a scalable framework for robust autonomy. Foutter’s contributions are particularly impactful for researchers in embodied AI, safe reinforcement learning, and human-robot interaction, as his methods reduce reliance on hand-coded safety constraints. With growing recognition for advancing foundation model applications in robotics, his research is shaping how future systems handle uncertainty—a key step toward trustworthy, general-purpose robots.