Varun Murali

I am currently a postdoctoral researcher at the University of Pennsylvania, where I work on hierarchical semantic mapping, planning, and tasking for robots operating in real-world environments. My research focuses on developing computationally efficient algorithms that enable robots to perform complex, under-specified tasks in dynamic and unstructured settings. I am particularly interested in designing general-purpose autonomy algorithms that integrate perception-action loops to improve decision-making and adaptability.

In past work, I have explored the synergistic connections between perception and planning for autonomous navigation, developing methods for reasoning about semantic affordances, improving robust feature matching for localization and mapping, and integrating perception-aware planning techniques to enhance robot navigation in challenging environments. Additionally, I am exploring how foundation models and large language models can be leveraged to create task representations that allow robots to understand and execute high-level instructions while maintaining situational awareness.

Looking ahead, I aim to advance research in hierarchical perception-action frameworks that enhance robotic decision-making and adaptability. I am particularly interested in the co-design of hardware and software solutions for size, weight, and power-constrained robots, as well as developing rule-based frameworks for safe and explainable robotic behavior.