Michael Posa
Assistant Professor, MEAM
Michael’s interests center on planning, control, and formal analysis of robots as they interact with the world. Whether a robot is assisting within the home, or operating in a manufacturing plant, the fundamental promise of robotics requires touching and affecting a complex environment in a safe and controlled fashion. Michael focuses on developing computationally tractable algorithms which enable robots to operate both dynamically and safely as they quickly maneuver through and interact with their environments. This research involves understanding the interplay between the non-smooth dynamics of contact and numerical optimization, and then testing these techniques on both legged robots and robotic arms.
Research Areas
- Contact Modeling
- Hybrid Systems
- Legged Locomotion
- Manipulation
- Optimal Control
- Autonomy
Research Groups
Projects
Publications
Push Anything: Single- and Multi-Object Pushing From First Sight with Contact-Implicit MPC
Publisher ICRA
Components of Contact: An Analysis of Contact-Implicit Control Strategies on Manipulation Primitives
Publisher ICRA
Publisher ArXiv
Object Reconstruction under Occlusion with Generative Priors and Contact-induced Constraints
Publisher IROS
Vysics: Object Reconstruction Under Occlusion by Fusing Vision and Contact-Rich Physics
Publisher RSS
Learning a Vision-Based Footstep Planner for Hierarchical Walking Control
Publisher IEEE-RAS International Conference on Humanoid Robots
Perceptive Mixed-Integer Footstep Control for Underactuated Bipedal Walking on Rough Terrain
Publisher IEEE Transactions on Robotics
Set-valued rigid-body dynamics for simultaneous, inelastic, frictional impacts
Publisher The International Journal of Robotics Research
Consensus Complementarity Control for Multicontact MPC
Publisher IEEE Transactions on Robotics
Dynamic on-palm manipulation via controlled sliding
Publisher Robotics: Science and Systems
Instance-Agnostic Geometry and Contact Dynamics Learning
Publisher IEEE
Simultaneous Learning of Contact and Continuous Dynamics
Publisher CoRL
Optimization-Based Control for Dynamic Legged Robots
Publisher IEEE
Validating Robotics Simulators on Real World Impacts
Publisher IEEE
Real-Time Multi-Contact Model Predictive Control via ADMM
Publisher ICRA
Learning Linear Complementarity Systems
Publisher L4DC
Publisher IEEE Robotics & Automation Magazine
Impact Invariant Control with Applications to Bipedal Locomotion
Publisher arXiv
ContactNets: Learning Discontinuous Contact Dynamics with Smooth, Implicit Representations
Publisher CoRL
Contact-Aware Controller Design for Complementarity Systems
Publisher arXiv
Modeling and Analysis of Non-unique Behaviors in Multiple Frictional Impacts
Publisher Robotics: Science and Systems (RSS)
Optimal reduced-order modeling of bipedal locomotion
Publisher arXiv
A Quasi-static Model and Simulation Approach for Pushing, Grasping, and Jamming
Publisher Workshop on the Algorithmic Foundations of Robotics (WAFR)
News
Faculty News
Touching the Future: Mastering physical contact with new algorithm for robots
October 10, 2024
Faculty News
Profs at Penn’s GRASP Lab snagged Toyota Research Institute funding for robotics research
February 1, 2021
Ph.D. Students
Alumni
Dr. Michael Posa and DAIR Lab
Research Areas
- Contact Modeling
- Hybrid Systems
- Legged Locomotion
- Manipulation
- Optimal Control
- Autonomy