New Energy and Industrial Technology Development Organization (NEDO) "Next-generation artificial intelligence and robot core technology development / Japan-US joint research and development of next-generation artificial intelligence technology / Assembly robot R & D based on HDR motion analysis technology"
In recent years, an increasing number of robots to achieve advanced assembly while adjusting force by equipped with force sensors. However, since the detection range (dynamic range) of the force sensor is narrow, it is not possible to measure minute forces, and it is difficult to adjust the force precisely. Therefore, in order to save labor in the industrial field, it is expected to develop and introduce assembly robots that use high-performance force sensors.
Therefore, since 2018, the New Energy and Industrial Technology Development Organization (NEDO) and Saitama University have started assembly robot research and development project based on High Dynamic Range (HDR) motion analysis technology for the purpose of automating high-order assembly operations. Among them, in cooperation with WACOH-TECH Inc., we developed an HDR force sensor that can detect a wide range of force from minute force to large force. In addition, in cooperation with KODENSHA CO.,LTD., we succeeded in developing the assembly robot used the developed HDR force sensor. The assembly robot equipped with an HDR force sensor is the first in the world.
This robot has a force detection range from 10g to 20kg, which is 10 times that of conventional commercial force sensors. Capable of detecting until 1 / 10th of the conventional force, it can assemble easily damaged objects such as resin materials with restrained force. In addition, the state of assembly can be accurately recognized by artificial intelligence (AI) technology using machine learning based on minute force information, enabling more advanced assembly work. For example, in the case of assembly with a click operation that generates vibration and clicking sound when finishing the assembly, improvements of the reliability of the work are achieved by the robot recognizes the success of the assembly by detecting the vibration as a force. This makes it possible to expand the robot's application range to work that requires delicate force adjustments in addition to the work that can be done with the conventional force sensor of the robot.
In the future, we plan to apply this technology especially to the assembly of resin materials and fragile materials, as well as complex assembly consisting of multiple procedures.
In addition, we will exhibit the newly developed assembly robot at the “the International Robot Exhibition 2019 (iREX2019)” to be held at Tokyo Big Sight from Dec. 18 to Dec. 21.
- Admittance Control Based on Stiffness Ellipse for Collision Force Control of Object Manipulation
- [in Japanese] HDR6軸力覚センサを用いたアドミッタンス制御
- Trajectory adjustment for nonprehensile manipulation using latent space of trained sequence-to-sequence model
- [in Japanese] クロスアーチ構造の導入によるHDR6軸力覚センサの小型化
- [in Japanese] ダイナミックな技術動作を学習するロボット
- [in Japanese] 前脛骨筋の張力を可視化するバイオフィードバック型筋力トレーニングシステム
- [in Japanese] 組み立て作業のための力制御における剛性楕円の自律生成
- [in Japanese] 微小な力感覚を認識するAI技術によってロボットの組み立ての信頼性向上