Max Robotics — Global Service Robot Directory + US Certification

About

Soshi Iba is a robotics researcher whose work spans human-robot interaction, robot learning, and manipulation of deformable objects. Early in his career, Iba made significant contributions to intuitive robot programming and control, developing gesture-based interfaces using hidden Markov models that allowed users to direct mobile robots through natural hand movements (61 citations) and pioneering multimodal programming systems that combined speech, gesture, and other intuitive inputs to make robots accessible to novice users (57 citations). More recently, Iba has emerged as a prominent voice in robotic fabric manipulation — a notoriously difficult problem due to complex deformable dynamics and high-dimensional state spaces. His team applied deep imitation learning to teach robots sequential fabric-smoothing policies from visual and depth data (109 citations), extended visual foresight frameworks to enable multi-task fabric manipulation (71 citations), and developed simulation-trained dense visual correspondences transferable to real-world folding and smoothing tasks (55 citations). His 2022 work on transfer learning for robot bed-making has already accumulated 81 citations, reflecting its immediate practical relevance. Rounding out his portfolio, Iba has addressed crowd navigation planning and visuotactile 6D pose estimation, demonstrating a broad research vision centered on enabling robots to operate skillfully and collaboratively in complex, human-centered environments.