Papers
41
Total Citations
784
H-Index
12
About
Maani Ghaffari is a robotics researcher whose work spans state estimation, motion planning, legged robot control, and multi-robot systems — areas collectively aimed at enabling robots to operate reliably and intelligently in complex, real-world environments. His most influential contribution, "Contact-aided Invariant Extended Kalman Filtering for Robot State Estimation" (2020, 295 citations), established a principled mathematical framework for legged robot pose and velocity estimation that overcomes the limitations of vision-dependent approaches, drawing on Lie group theory to achieve robust, geometry-aware filtering. This geometric perspective recurs throughout his research, including error-state Model Predictive Control on matrix Lie groups and invariant Kalman filtering for slip-aware mobile robot navigation. Beyond state estimation, Ghaffari has made notable strides in informative motion planning, developing sampling-based algorithms with information-theoretic guarantees for robotic exploration and environmental monitoring. His work on legged robot terrain traversability using deep inverse reinforcement learning and safety-aware planning with Control Barrier Functions reflects a commitment to deployable, safety-conscious autonomy. He has also contributed to heterogeneous multi-robot task scheduling under uncertainty and multilayer Bayesian mapping. With over 600 combined citations across his top papers, Ghaffari has emerged as a distinctive voice bridging mathematical rigor and practical robotics.
Research Focus
Key Achievements
Top Papers
- 1Contact-aided invariant extended Kalman filtering for robot state estimation295 citations · 2020
- 2Model-based PI–fuzzy control of four-wheeled omni-directional mobile robots66 citations · 2011
- 3
- 4
- 5
- 6
- 7
- 8Toward Safety-Aware Informative Motion Planning for Legged Robots22 citations · 2021
- 9
- 10Multitask Learning for Scalable and Dense Multilayer Bayesian Map Inference17 citations · 2022