Self-Resetting Soft Ring Robot Achieves Continuous Jumping via Physical Intelligence

Miniature soft jumping robots have long faced an engineering bottleneck: they can perform a single efficient jump, but afterwards typically require manual reset, mechanical latches, or switching external stimuli to initiate the next jump, making continuous motion under unsupervised conditions difficult. A team led by Professor Jie Yin at North Carolina State University, publishing in PNAS, has coupled geometric structure with photoresponsive materials to achieve autonomous, uninterrupted jumping under constant infrared illumination, opening a new path for miniature soft robot locomotion.

Dubbed the ring jumper, the robot's main body is a teardrop-shaped band made of liquid crystal elastomer, with a V-shaped rigid aluminum tail tube attached at one end. When continuously illuminated by infrared light, the material undergoes photothermal contraction, driving the band to rotate. The rigid V-shaped tail prevents rolling in place, forcing the band to twist and accumulate elastic potential energy. When deformation reaches a critical threshold, buckling instability triggers a rapid energy release, causing the tail to strike the ground and launch the robot into the air. During flight, the soft ring untwists and returns to its original form, completing a self-reset. As long as the infrared source remains on, the cycle of energy storage, release, reset, and takeoff repeats continuously, with no onboard electronics controlling the process.

Geometric parameters serve as the primary switch for regulating the robot's motion modes. Research data show that the angle of the V-shaped tail directly determines behavior: a 120-degree angle results in ground crawling, a 90-degree angle enables directional forward jumping, and a 50-degree angle switches to vertical jumping. Researchers can also add weights at the end of the ring to shift the center of mass, optimizing forward jumping force and stability. Infrared light intensity affects performance as well: a minimum threshold is needed to trigger jumping, but excessive intensity causes chaotic trajectories and loss of predictable direction.

Prototype tests verified adaptability to diverse unstructured environments. The robot can climb slopes, clear low obstacles, and maintain continuous jumping on grass, sand, gravel, humus, and water-land interfaces. In optimal configurations, it achieves vertical jumps up to 80 times its own size and horizontal distances over three times its body length. Thanks to the structure's impact tolerance, the robot resumes its motion cycle after collisions—a key feature for miniature robots in complex field environments.

The core innovation lies in physical intelligence: encoding locomotion logic into material properties and geometric configuration rather than relying on sensors, chips, and real-time algorithms. Traditional miniature robots aiming for continuous jumping often require driving circuits and sensing modules, adding weight, size, and power consumption, which complicates design. This approach delegates control logic to material deformation and mechanical instability, simplifying the overall structure and reducing hardware complexity. However, it has inherent limitations: motion modes are predefined, and the robot cannot change its movement strategy in real time; only pre-design adjustments to geometry and mass distribution can alter behavior.

In terms of industry status, the current work remains at the proof-of-concept stage, with several issues to address before real-world application. The robot depends on external infrared light; without such illumination in field environments, the drive mechanism fails. Repeated large deformations of the liquid crystal elastomer raise concerns about material fatigue and aging, requiring long-term testing. Additionally, lacking sensing and communication modules, the robot can only demonstrate motion and cannot collect or transmit environmental data.

Despite these challenges, the research holds significant reference value. With iterative optimization, such low-cost miniature soft robots have potential to develop into swarm robots. Large numbers of tiny robots could be mass-deployed to traverse rugged terrain via jumping for environmental surveys or preliminary disaster site inspections. Corresponding author Jie Yin stated that there is no immediate commercial product, but this fundamental research can provide new ideas for environmental navigation, swarm robotics, and unstructured terrain detection. The study was conducted by Fangjie Qi, Caizhi Zhou, Haitao Qing, Haoze Sun, Yaoye Hong, and Jie Yin, with support from several grants from the U.S. National Science Foundation. The work also offers a lesson for the soft robotics field: beyond advancing algorithms and control systems, fully exploiting physical intelligence from materials and geometric structures is a direction that cannot be ignored for solving miniaturization and low-power challenges.