Shahram Khorshidi

About

Shahram Khorshidi is pushing the boundaries of what legged robots can achieve, specializing in state estimation and perception for dynamic locomotion. His work tackles the core challenge of enabling robots to move with agility and precision, even during complex maneuvers like flight phases. Khorshidi’s major contributions include fusing visual-inertial data with leg odometry to maintain reliable posture estimation when traditional methods fail, and pioneering the use of Koopman operator theory to linearize nonlinear dynamics for centroidal state estimation—a breakthrough that enhances predictive control. He has also explored the novel use of torque measurement in state estimation, a sensor modality often overlooked, and developed methods for physically-consistent parameter identification of robots in contact. With his most-cited paper, "Perception for Humanoid Robots," garnering 24 citations, Khorshidi’s impact is evident in advancing the safety and efficiency of human-robot interaction. His work, published in top venues from 2022 to 2025, is essential reading for anyone interested in the future of agile, perceptive robots.

Research Focus

Key Achievements

Top Papers

1. 1
   Perception for Humanoid Robots

   24 citations · 2023
2. 2
   Visual-Inertial and Leg Odometry Fusion for Dynamic Locomotion

   9 citations · 2023
3. 3
   Centroidal State Estimation Based on the Koopman Embedding for Dynamic Legged Locomotion

   4 citations · 2024
4. 4
   On the Use of Torque Measurement in Centroidal State Estimation

   4 citations · 2023
5. 5
   Physically-Consistent Parameter Identification of Robots in Contact

   2 citations · 2025
6. 6
   Visual-Inertial and Leg Odometry Fusion for Dynamic Locomotion

   2 citations · 2022

Key Collaborators

Contact & Links