AS Soembagijo

About

A. S. Soembagijo is a pioneering researcher in intelligent robotic manipulation, with a focus on sensor-based control and machine learning for assembly tasks. His work bridges the gap between traditional robotics and adaptive learning systems, particularly in contact-rich operations. Soembagijo’s most influential contribution, “Featureless classification of tactile contacts in a gripper using neural networks” (1997, 33 citations), introduced a novel approach to interpreting tactile sensor data without explicit feature extraction, enabling robots to classify contact states directly from raw force signals. This work laid the groundwork for more robust and flexible gripper control in unstructured environments. Earlier, in “Learning the peg-into-hole assembly operation with a connectionist reinforcement technique” (1995, 6 citations), Soembagijo developed a learning controller that autonomously increased insertion speed during consecutive peg-into-hole operations while maintaining low contact forces. By using reinforcement learning to optimize the relationship between measured forces and controlled velocity, he demonstrated how robots could improve assembly performance through experience, reducing the need for complex analytical models. Though his citation counts reflect a focused niche, Soembagijo’s research remains foundational for those exploring neural network applications in tactile sensing and adaptive assembly automation.

Research Focus

Key Achievements

Top Papers

1. 1
   Featureless classification of tactile contacts in a gripper using neural networks

   33 citations · 1997
2. 2
   Learning the peg-into-hole assembly operation with a connectionist reinforcement technique

   6 citations · 1995

Key Collaborators

Contact & Links