From cleaning up plastic waste to recovering valuable resources and exploring seabed ecosystems, underwater operations demand grippers that are both gentle and strong. Traditional rigid grippers lack adaptability, while soft ones struggle with heavy loads. Now, an international research team led by Professor Xie Guangming at Peking University has developed an octopus-inspired solution that overcomes this trade-off—delivering the fastest stiffness-switching performance reported to date.
The new device, called the Octopus-Inspired Upward Transport Robot (OUT-Robot), mimics the multimodal grasping strategy of an octopus. Its secret lies in a specially designed variable-stiffness arm that uses a shape memory polymer (SMP)—polylactic acid (PLA)—combined with a trilayer thermal interface. "We engineered a thermodynamic synergy among the material, geometry, and the underwater environment," explains Professor Xie. "The inner silicone layer diffuses heat uniformly, the outer layer acts as a transient barrier during heating, and the surrounding water becomes an active heat sink during cooling."
This clever design delivers stunning speed: the gripper softens in just 1.3 seconds when voltage is applied and rigidifies in 0.8 seconds once heating stops. "Our stiffness transition time is substantially faster than any previously reported actuator," Xie adds. In contrast, conventional SMP-based systems often take tens of seconds to cool down in air.
The key innovation is a "Soft-Rigid Hybrid" manipulation approach. During grasping, the arms remain compliant, allowing the integrated suckers to conform perfectly to irregular surfaces. Once the object is secured, the SMP is rapidly cooled and locks the shape—holding heavy items without any continuous energy input. "This zero-energy shape locking is a game changer for long-duration underwater missions," says Professor Xie.
Experimental results confirm the dramatic improvement. A single arm with stiffened SMP shows roughly 25 times higher stiffness than an arm without SMP. The six-arm gripper achieves cooperative grasping forces exceeding 4 newtons (equivalent to >400 grams) in Mode VI (six-arm grasp). Suction mode adds further versatility, with pre-adhesion force increased 2.2-fold compared to previous designs.
To prove its practical value, the team tested the gripper in a 2-meter-deep pool cluttered with stones, fishing nets, plastic bottles, sea cucumbers, scallops, a fragile plate, an aluminum profile, and a heavy 500-gram beer bottle. The gripper seamlessly switched between different grasping modes—removing a light fishing net (<1 g), collecting delicate biological samples, and hoisting the heavy bottle. "Our robot can handle objects from extremely light debris to heavy solid waste over 500 grams, all in one continuous operation," Xie highlights.
The OUT-Robot integrates active buoyancy control with manipulation. After grasping an object and locking the SMP, the robot inflates its soft shell to increase buoyancy and floats upward—transporting the object to the surface passively. "The grasping phase consumes about 75 joules for 1.3 seconds, while the subsequent ascent uses almost zero energy," notes Professor Xie. This "pulse-actuation, zero-retention" strategy dramatically reduces total energy consumption compared to continuously powered systems. Additionally, the robot can crawl omnidirectionally underwater by coordinating arm bending, achieving 70 cm in 55 seconds along a fixed direction.
The authors believe this technology opens new avenues for autonomous marine operations, ecological restoration, and ocean missions. "By turning the underwater environment from a thermal burden into an active component of the control loop, our design is scalable and could be deployed in swarms for distributed collection tasks," says Professor Xie. "We are providing a robust, efficient, and quiet solution to protect our oceans—one grasp at a time."
Authors of the paper include Mingxin Wu, Yurong Liu, Jiaxi Wu, Waqar Hussain Afridi, Xingwen Zheng, Chen Wang, and Guangming Xie.
This work was supported by the National Natural Science Foundation of China (grants U23B2037, U22A2062, and 12272008), the Beijing Natural Science Foundation (grant 3242003), the Postdoctoral Fellowship Program of CPSF (grants GZB20250072 and 2025M781342), and the Key Technology Research and Development Program of Henan Province (grant 252102221024).
The paper, "Octopus-Inspired Underwater Gripper with Rapid Stiffness Tuning and Robot Enabling Upward Transport" was published in the journal Cyborg and Bionic Systems on Mar. 31, 2026, at DOI: 10.34133/cbsystems.0528.
