Max Robotics — Global Service Robot Directory + US Certification

About

Arsalan Mousavian is a robotics researcher whose work sits at the intersection of computer vision, robot manipulation, and embodied AI. His research spans several interconnected domains: 6D object pose estimation and tracking, robotic grasping, scene perception, and the application of large language models to robot task planning. Mousavian has made significant contributions to enabling robots to understand and interact with their physical environments. His ProgPrompt framework (508 citations) demonstrated how large language models could generate structured, executable task plans for situated robots—a landmark contribution to the LLM-for-robotics movement. His work on PoseRBPF introduced a principled probabilistic approach to 6D object pose tracking, while his self-supervised pose estimation system addressed the costly challenge of real-world data annotation. Tackling a persistent perception gap, he developed methods for segmenting unseen objects and recovering depth information for transparent surfaces—both critical for practical robot deployment. His systematic review of deep learning approaches to grasp synthesis (215 citations) has become a valuable reference for the robotics community, and his human-to-robot handover research advances safe, reactive physical human-robot interaction. With over 1,600 cumulative citations across these contributions, Mousavian has established himself as a leading figure in perception-driven robot manipulation research.