One of the most promising areas for applying manipulation research is industrial robotics. In fact, automation of parts supply, product assembly and product inspections processes are considered extremely difficult. The labor-intensive nature of these processes has contributed to the hollowing out of domestic manufacturing industry due to rise of labor costs. In particular, the automation of the parts supply and product assembly requires the full use of manipulation research.

Parts Supply Operation

First, we will focus on the parts supply process. When considering the automation of product assembly, the question is how to automate the supply of parts to the assembly process. Automation becomes especially difficult when the number of parts that make up the product increases. The parts that make up the product may have a variety of shapes and include soft parts such as rubber. These various parts must be accurately recognized by visual sensors and accurately grasped by hand, which is exactly what is required in research on grasping by robot hand.


In addition, e-commerce has been widely used in recent years, and when a product order is received at an e-commerce distribution center, a person searches for the ordered product, picks up the product, packs it in a box and ships it to the consumer. The process of picking up and packing these products by human operators presents almost the same problem as the automation of the parts supply process described above, and this automation is considered a very important issue.


Assembly Work

Next, we consider the actual assembly of products by robots. In recent years, in order to meet the consumer's demands, high variety productions has been widely introduces as a form of product manufacturing. Currently, it is no longer sufficient for robots to perform only routine and repetitive tasks as in the past. Specifically, the following problems must be solved:

  • Teaching cost: It is difficult for a person to teach the robot to perform a certain operation, so it is necessary to consider how to facilitate the teaching of the operation when the manufacturing process changes frequently in high variable-low volume production.
  • Part grasping: when performing assembly operations, the hand must securely grip the part. In this case, specialized hands are often used that can grip specific parts reliably. In this case, it is necessary to change hands using a device called a tool changer each time the part to be grasped changes. The number of hands that must be prepared beforehand increases as the number of parts increases.
  • Existence of difficult processes: If a robot is to replace humans in tasks that are performed based on previous experience and knowledge of humans, such as inserting a flexible cable, it is necessary to extract the specific knowledge that humans use to perform such tasks, and apply them to the robot.

Research Topics

Robotic Picking research


  • Kensuke Harada, Weiwei Wan, Tokuo Tsuji, Kohei Kikuchi, Kazuyuki Nagata, and Hiromu Onda: Initial Experiments on Learning-Based Randomized Bin-Picking Allowing Finger Contact with Neighboring Objects, Proc. of IEEE Int. Conf. on Automation Science and Engineering, pp. 1196-1202, Fort Worth, USA, August, 2016.
  • Kensuke Harada, Weiwei Wan, Tokuo Tsuji, Kohei Kikuchi, Kazuyuki Nagata, and Hiromu Onda: Iterative Visual Recognition for Learning Based Randomized Bin-Picking, Preprints of 2016 Int. Symposium on Experimental Robotics, Tokyo, Japan, October, 2016.
  • Kensuke Harada, Weiwei Wan, Kohei Kikuchi, Kazuyuki Nagata and Hiromu Onda, “Experiments on Learning Based Industrial Bin-picking with Iterative Visual Recognition,” Industrial Robot: an International Journal, vol. 45, no. 4, pp. 446-457, DOI 10.1108/IR-01-2018-0013, 2018.
  • Ryo Matsumura, Kensuke Harada, Yukiyasu Domae, and Weiwei Wan, “Learning Based Industrial Bin-picking Trained with Approximate Physics Simulator,” Proceedings of International Conference on Intelligent Autonomous Systems, #33, 2018.
  • Ryo Matsumura, Yukiyasu Domae, Weiwei Wan, and Kensuke Harada, “Learning Based Robotic Bin-picking for Potentially Tangled Objects,” Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 7984-7991, 2019.
  • Yuya Sato, Kensuke Harada, Nobuchika Sakata, Weiwei Wan, and Ixchel G. Ramirez-Alpizar, “Two-stage Picking Method for Piled Shiny Objects,” Proceedings of the 15th IFToMM World Congress on Mechanism and Machine Science, (T.Uhl(ed.), Advances in Mechanism and Machine Science, Mechanism and Machine Science 73, Springer), pp. 2049-2058, 2019.
  • Kenta Matsuura, Keisuke Koyama, Weiwei Wan, and Kensuke Harada, “Robotic Picking for Piled Sushi Topping,” Proceedings of the 2021 International Conference on Artificial Life and Robotics, pp. 328-331, 2021.
  • Xinyi Zhang, Keisuke Koyama, Yukiyasu Domae, Weiwei Wan, Kensuke Harada, "A Topological Solution of Entanglement for Complex-shaped Parts in Robotic Bin-picking," Proceedings of IEEE International Conference on Automation, Science and Engineering, pp. 461-467, 2021.