I’m a PhD student in Robotics at the ARM lab at the University of Michigan, advised by Dmitry Berenson. I’m interested in utilizing planning, control theory and learning for robotic manipulation.

Publications

• Thomas Power and Dmitry Berenson. Variational Inference MPC using Normalizing Flows and Out-of-Distribution Projection. Robotics: Science and Systems, June 2022. arXiv

• Thomas Power and Dmitry Berenson. Improving Sample-based MPC with Normalizing Flows & Out-of-distribution Projection. ICRA Workshop on Motion Planning with Implicit Representations of Geometry, May 2022.

• Thomas Power and Dmitry Berenson. Variational Inference MPC for Robot Motion with Normalizing Flows. NeurIPS Robot Learning Workshop, December 2021.

• Thomas Power and Dmitry Berenson. Keep it Simple: Data-efficient Learning for Controlling Complex Systems with Simple Models. IEEE Robotics and Automation Letters (RA-L), April 2021. arXiv

• Dale McConachie, Thomas Power, Peter Mitrano, and Dmitry Berenson. Learning When to Trust a Dynamics Model for Planning in Reduced State Spaces. IEEE Robotics and Automation Letters (RA-L), April 2020. arXiv