Khashayar Rohanimanesh
Michigan State University, University of Massachusetts Amherst
Papers
6
Total Citations
85
H-Index
4
About
Khashayar Rohanimanesh is a leading researcher in robotics and artificial intelligence, focusing on hierarchical decision-making, robot navigation, and learning from demonstrations. His foundational work on hierarchical partially observable Markov decision processes (POMDPs) for robot navigation, published in 2001, has garnered over 50 citations and established frameworks for modeling complex, partially observable environments like office buildings. His 2002 follow-up paper further advanced hierarchical hidden Markov models (HHMMs) for efficient robot navigation. Rohanimanesh also made significant contributions to imitation learning, developing methods to learn from visual data with multiple intentions, enabling robots to acquire complex skills from raw image inputs. His research on concurrent decision-making in Markov decision processes, detailed in his 2006 dissertation, addresses fundamental coordination challenges in multi-action systems. More recently, he has explored self-supervised learning for goal-conditioned pick-and-place tasks and online tool selection using learned grasp prediction models, demonstrating practical applications in robotic bin-picking systems. With over 85 total citations across his most-cited works, Rohanimanesh’s research continues to influence both theoretical frameworks and real-world robotic systems, bridging hierarchical modeling, imitation learning, and autonomous manipulation.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Imitation Learning from Visual Data with Multiple Intentions6 citations · 2018
- 4Concurrent decision making in markov decision processes6 citations · 2006
- 5Self-Supervised Goal-Conditioned Pick and Place4 citations · 2020
- 6Online Tool Selection with Learned Grasp Prediction Models3 citations · 2023