Edgar Anarossi

Nara Institute of Science and Technology

Papers

Total Citations

H-Index

About

Edgar Anarossi is a robotics researcher whose work sits at the intersection of machine learning, control theory, and soft robotics. His primary research areas include data-driven modeling for complex dynamical systems, imitation learning for human-robot interaction, and motion representation for discontinuous tasks. Anarossi’s most notable contribution is his work on Deep Koopman with Control, where he applies spectral analysis to soft robot dynamics—a notoriously difficult domain due to inherent non-linearities like elasticity and deformation. This paper, with 11 citations, offers a powerful alternative to explicit physics-based modeling by learning control models directly from data. In the realm of human-robot collaboration, his work on Disturbance Injection Under Partial Automation (5 citations) advances robust imitation learning for long-horizon tasks, addressing the critical challenge of maintaining performance under real-world perturbations. Additionally, his research on Deep Segmented DMP Networks (4 citations) tackles the underexplored problem of learning discontinuous motions—essential for state-aware robotic tasks that require sudden changes in direction or velocity. Anarossi’s work is characterized by its practical focus on enabling robots to handle the messy, non-ideal conditions of real-world operation, making him a rising voice in the field of learning-based robot control.