AGIBOT's 10,000th Humanoid, HEAPGrasp Goes Mainstream, and the Factory Floor Gets Smarter

Research

HEAPGrasp Hits 96% Grasp Success on Transparent and Reflective Objects Using Only RGB Camera

Tokyo University of Science's HEAPGrasp uses shape-from-silhouette and semantic segmentation to reliably handle glass and shiny metal without depth sensors, cutting camera travel by 52% and execution time by 19%. For anyone building or deploying pick-and-place automation in logistics or food service, this is a practical drop-in upgrade path since it retrofits onto existing robotic arms.

HEAPGrasp Paper: Full Technical Details on Hand-Eye Active Perception for Diverse Optical Properties

The IEEE Robotics and Automation Letters publication confirms a 96% grasp success rate across transparent, specular, and opaque objects with a trajectory length reduction of 52%. Worth bookmarking for teams evaluating vision stack upgrades or building grasp planning pipelines from scratch.

News

Japanese Vision System Cracks the Transparent Object Problem for Industrial Robots

The core insight here is that silhouette-based 3D reconstruction sidesteps the depth-sensor failure mode entirely, which has been a persistent bottleneck in automated handling lines. A 96% success rate across opaque, transparent, and specular objects in real-world tests is the kind of number that moves procurement conversations.

Analysis

RGB-Only Robotic Grasping Now Works on Glass and Shiny Metal in Mixed Industrial Scenes

HEAPGrasp's deep learning-based next-pose planner is the underrated piece here since it actively minimizes camera movement rather than just brute-forcing more viewpoints. That efficiency gain is what makes it viable on a real production timeline, not just a lab demo.

Events

Tokyo University of Science Publishes HEAPGrasp Method Headed to ICRA 2026

The ICRA 2026 presentation slot signals this work is getting serious peer validation, not just a journal drop. If you're tracking the sim-to-real and perception stacks being standardized for industrial arms, this is one to watch heading into the conference season.