Ankit Shah

Papers

Total Citations

H-Index

About

Ankit Shah is a leading researcher in the intersection of robotics, formal methods, and human-robot interaction, with a focus on ensuring both safety and efficiency in autonomous systems. His work addresses critical challenges in motion planning and reward specification, particularly when robots must operate alongside humans or interpret complex, time-dependent tasks. Shah’s most cited paper, “Provably Safe and Efficient Motion Planning with Uncertain Human Dynamics” (32 citations), introduces a formal framework that guarantees human physical safety—defined as collision avoidance or safe impact—without sacrificing task performance during interactive manipulation. He further advanced the field with “Planning With Uncertain Specifications (PUnS)” (16 citations), which tackles the difficult problem of reward engineering by enabling robots to handle non-Markovian reward structures expressed in linear temporal logic. In his interactive robot training research, Shah developed Bayesian frameworks that allow robots to learn temporal and non-Markov tasks from human demonstrations without over-constraining policies, making it easier for non-experts to teach robots complex behaviors. His work is foundational for deploying robots in homes and workplaces, where they must understand nuanced human instructions while maintaining rigorous safety guarantees.