OMRON SINIC X Showcases Robotics Innovations at IROS 2025
OMRON SINIC X Corporation (OSX), a leading-edge robotics research arm of OMRON, is making headlines at the 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2025), one of the world’s top-tier robotics events. Held in Hangzhou, China from October 19 to 25, IROS 2025 is celebrating an impressive acceptance rate, with 1,991 papers—about 46.2%—selected from over 4,300 global submissions. This year, OSX will unveil two pioneering research projects that could shape the future of intelligent manipulation in unstructured environments.
Research Spotlight 1: Cable-driven Serpentine Manipulators
Unconventional Design for Lightweight, Dexterous Robots
The first paper from OSX, led by Kazutoshi Tanaka and colleagues, introduces a new class of cable-driven serpentine manipulator arms. These manipulators offer unique advantages:
Exceptional obstacle avoidance and multi-directional force application
Lightweight design by using plastic links and consolidating all motors and sensors at the base
Their prototype features a 9-degree-of-freedom arm, stretching 545 mm and weighing only 308 g—much lighter than conventional robotic arms. However, this ultralight configuration poses major control challenges due to:
Cable slack and elongation
Link deformation
Resulting discrepancies between predicted and actual arm positions
AI-powered Pose Estimation: Physical Reservoir Computing
To overcome these challenges, OSX researchers propose a novel pose estimation method leveraging "physical reservoir computing." Instead of treating the manipulator's inherent nonlinearities as a problem, they exploit this behavior as a computational resource. Experimental results show remarkable accuracy, with a mean pose error of just 4.3 mm—much better than a conventional analytical approach (39.5 mm error) and on par with advanced neural networks like LSTM (4.4 mm error).
Key Benefits:
Enhanced accuracy in controlling lightweight, flexible arms
Simpler and more efficient than relying solely on deep learning or pure physical models
Research Spotlight 2: Variable Stiffness Actuators with WAVE
Introducing WAVE: Worm Gear-based Adaptive Variable Elasticity
The second significant research, by Moses Gladson Selvamuthu and team, focuses on increasing robotic safety and adaptability through a new actuator design:
WAVE incorporates a non-backdrivable worm gear with an adaptive elastic element (spring)
Decouples the driving motor from external forces, ensuring that external shocks do not directly stress the motor
WAVE lets operators precisely control the joint stiffness by altering the spring’s compression, absorbing shocks and protecting the actuator from excessive loads. The stiffness can be continuously tuned—even during operation.
Applications and Impact
The actuator’s ability to store impact forces as elastic energy and maintain load even when the motor is idle (at rest, under load) makes it ideal for tasks involving frequent human-robot interaction or unpredictable contacts.
Robust force transmission and compliance enable longer and safer robot operation in challenging, contact-rich settings, such as industrial automation, disaster response, or medical robotics.
Why This Matters
Enabling the Next Generation of Safe, Intelligent Robots
Both innovations exemplify the fusion of artificial intelligence, advanced mechanics, and novel sensing for next-generation robotics. OSX’s work signals a shift towards:
More adaptable and resilient robots
Intelligent exploitation of physical system nonlinearities
Increased safety in environments where robots and humans collaborate
About OMRON SINIC X
OSX is a strategic subsidiary of OMRON, dedicated to forecasting and realizing “near-future design.” Their interdisciplinary team collaborates with universities and institutes worldwide, driving advancements in AI, robotics, IoT, and sensing to address societal challenges and accelerate innovation.