Untethered Miniature Tensegrity Robot with Tunable Stiffness for High-Speed and Adaptive Locomotion

Abstract

Miniature robots are increasingly used in unstructured environments and require higher mobility, robustness, and multifunctionality. However, existing purely soft and rigid designs suffer from inherent defects, such as low load capacity and compliance, respectively, restricting their functionality and performance. Here, we report new soft-rigid hybrid miniature robots applying the tensegrity principle, inspired by biological organisms’ remarkable multifunctionality through tensegrity micro-structures. The miniature robot’s speed of 25.07 body lengths per second is advanced among published miniature robots and tensegrity robots. The design versatility is demonstrated by constructing three bio-inspired robots using miniature tensegrity joints. Due to its internal load-transfer mechanisms, the robot has self-adaptability, deformability, and high impact resistance (withstand dynamic load 143,868 times the robot weight), enabling the robot to navigate diverse barriers, pipelines, and channels. The robot can vary its stiffness to greatly improve load capacity and motion performance. We further demonstrate the potential biomedical applications, such as drug delivery, impurity removal, and remote heating achieved by integrating metal into the robot.

Introduction

Miniature robots, owing to their compact size, are widely used for navigation and exploration tasks in confined spaces. For example, they can efficiently navigate the small pipelines (ranging from a few millimeters to several centimeters) used for transporting water, gas, and oil inside aircraft engines and refinery machinery, enabling timely detection of potential defects.¹ Miniature robots can also enter internal cavities through the body’s natural passages and ducts in a minimally invasive manner, supporting critical procedures such as biopsies and targeted drug delivery.^2–5 Generally, miniature robots are categorized into two types: rigid and soft. Miniature rigid robots have precise motion;⁶ however, due to the limitations of their rigid structures, they exhibit limited compliance in unstructured environments. In contrast, miniature soft robots, composed of soft materials, display strong adaptability and multimodal motion capabilities^7,8; though, they are limited in load-carrying capacity and structural stiffness due to the materials constraints. Additionally, the soft bodies hinder the integration of nonmagnetic external stimuli,⁹ restricting remote heating capability for biomedical applications. This is because localized heat generation requires solid metallic materials for increased efficiency. In a word, purely miniature soft and rigid robots have some inherent defects, respectively, and face the inherent trade-off between working capability and compliance.

Reflecting on nature’s remarkable solutions to such challenges, we observed most biological organisms, such as mammals, rely on multiscale general tensegrity structures,¹⁰ which are soft-rigid hybrid, to achieve remarkable multifunctionality, such as fast locomotion, stiffness variation, high load-bearing capacity, and excellent environmental adaptability.^11,12 The tensegrity structures can be found in the cytoskeleton,¹³ cell membranes, DNA structures,¹⁴ and the skeletal muscle system in mammals, including the joints, limbs, and spine.^15,16 For example, the cytoskeleton utilizes tensegrity’s tension balance to maintain cell shape and mechanical stability, enabling the cell to interact with its environment and perform various movements.¹³ General tensegrity structures¹⁷ are composed of tension elements, such as cables and springs, or muscles in animals, and compression elements, such as straight, curved bars,¹⁸ or complex rigid bodies.^19–21 It can achieve a coordinated balance between flexibility and rigidity, and may address the challenges of purely miniature soft and rigid robots,^10,22 with great potential for constructing high-performance miniature robots.

This feasibility is further demonstrated by the extensive development of tensegrity robots,^23,24 which are capable of performing various motions, including crawling,²⁵ rolling,²⁶ flying,²⁷ variable-stiffness swimming,^19,21,28 and multistable grasping.²⁹ The robots also show significant potential for applications in areas such as space exploration, as proposed by NASA,²⁶ pipeline navigation,^30,31 and bionic design.³² However, as shown in Figure 6d in this article, existing tensegrity robots are typically over about 10 cm in size and have relatively slow speeds, which limits their applications in complex and confined environments. For example, the tensegrity robotic fish driven by a single motor has a body length of 36 cm and a maximum swimming speed of 0.87 body length/second (BL/s).¹⁹ The tensegrity mobile robot driven by six motors has a body length of 33 cm and a maximum speed of approximately 19.9 cm/s (0.58 BL/s).³³

In addition, soft-rigid coupled miniature robots encounter notable challenges in assembly.⁴⁶ Miniature tensegrity structures exhibit unique structural characteristics such as complex spatial topology,¹⁰ static and kinematic indeterminacy,^47,48 high sensitivity to prestress,^49,50 and a coordinated yet antagonistic tension network.¹⁷ These characteristics increase assembly complexity, presenting difficulties such as controlling the prestress of independent tension elements and precisely constructing a complex three-dimensional (3D) tension network in space. A rapid assembly method for tensegrity structures has been proposed using modular elastic lattices⁵¹; however, it has not yet been applied to the design of miniature tensegrity structures. Tensegrity structures can also be directly manufactured using 3D printing combined with sacrificial molding methods.⁵² However, this approach does not provide pretension within the elastic network. Actually, pretension is essential for ensuring stiffness and stability in tensegrity structures.¹⁷

In this work, we designed and implemented various types of miniature tensegrity robots with sizes between 14 and 20 mm, after addressing miniature tensegrity structure’s assembly challenges. First, we present approaches for the fabrication and actuation of a miniature tensegrity robot, using a miniature spherical tensegrity robot as a case study. Next, miniature prismatic and rotational tensegrity joints are designed to construct tensegrity crawlers and swimmers, highlighting the broad design possibilities of such robots. Then, we demonstrate the multimodal motion capabilities, high compressibility, and strong environmental adaptability of miniature tensegrity robots. The effectiveness of tunable stiffness in the miniature tensegrity swimmer is validated, showing its significant influence on swimming performance in intermediate flow regimes. Finally, the potential biomedical applications of miniature tensegrity robots are demonstrated, including high-speed drug delivery, impurity removal, and remote heating.

Results

Cell-inspired miniature spherical tensegrity robot with spatial antagonistic tension network

Inspired by that white blood cells rely on a spherical tensegrity cytoskeleton to roll along blood vessel walls,^13,53 we developed a miniature spherical tensegrity robot. Its assembly method with six steps can address the assembly challenge (see Supplementary Section 1.1 in Supplementary Data) and be found in Supplementary Section 1.2 in Supplementary Data.

As shown in Figure 1a, the robot’s spherical structure consists of six bars, 12 nodes, and 24 tension elements. The lengths of the tension elements and bars were set to 9.798 and 15 mm, respectively. The bars and 3D tension network are separated, as shown in Figure 1b. Next, the spatial tension network is planarized, converting it into an integrated planar elastic network in Figure 1c. The integrated planar elastic network has two repeated nodes, n₁₀ and n₁₂, created after planarization. By applying this method, the individual and independent tension elements of the spatial tension network are integrated into a single unit, thereby reducing assembly steps, lowering complexity, and shortening the overall assembly time. The remaining assembly steps can be found in Supplementary Section 1.2 in Supplementary Data.

FIG. 1.

Miniature spherical tensegrity robot. (a) Miniature spherical tensegrity structure. (b) Composition of the miniature spherical tensegrity structure includes bars and a spatial tension network. (c) Planar elastic network. (d) The 3D model and physical prototype of the robot. (e) Rolling on the ground by a rotating magnetic field. (f) Deformation under the influence of a permanent magnet. (g) Experimental images of deformation under the 2 kg load. (h) Experimental images of deformation under dynamic compression by a 54 kg electric vehicle with an adult weighing 68 kg.

Magnetic actuation is employed to remotely manipulate the motions of the miniature tensegrity robots. This actuation enables the separation of the power source and control system from the robot, allowing the robot to move freely without constraints. In Figure 1d, the magnetic plates are attached to two parallel bars, respectively, with volume V and in-plane magnetization M. The manufacturing method of the magnetic plates is detailed in Supplementary Section 2.3 in Supplementary Data. Without the actuation, the robot remains in a stable and self-balanced state through its tension network. The two magnetized plates align with the initial direction of the rotating magnetic field B. Driven by the rotating magnetic field B, the magnetic plate generates the torque T = V(M × B), causing the bars to rotate and align with the magnetic field direction. Meanwhile, the movement of the bars pulls the adjacent tension elements, causing the entire robot to rotate. In Figure 1e and Supplementary Movie S1, the robot can roll forward through elastic interactions with the environment, similar to the movement of spherical white blood cells.¹³

The miniature spherical tensegrity robot also exhibits magnetically controlled deformation capabilities. In Figure 1f and Supplementary Movie S2, under the influence of a permanent magnet, the robot actively compresses by one-third. To further demonstrate its compressibility, by applying 200 and 2000 g weights, the robot can deform significantly, compressing to 10% of its original volume. After load removal, the robot returns to its original shape, with no observed fractures or irreversible deformation, as shown in Figure 1g. The robot weighs approximately 0.424 g and can withstand static loads up to 4716 times its weight. Under the dynamic compression from a 54 kg electric vehicle with an adult weighing 68 Kg, the robot returns to its original shape and jumps, in Figure 1h and Supplementary Movie S3. The miniature tensegrity spherical robot possesses high compressibility and elastic recovery capabilities, enabling it to withstand dynamic loads up to approximately 143,868 times its weight, assuming both tires bear the same weight.

Miniature tensegrity joints and their robotic applications

In order to show the design versatility, we proposed miniature prismatic and rotational tensegrity joints that enable the flexible design and construction of miniature tensegrity robots capable of performing various motion modes. The prismatic tensegrity joint has a translational degree of freedom (DOF) and is composed of a planar D-bar structure,¹⁰ which is classified as a Class 2 structure. As illustrated in Figure 2a, the prismatic joint consists of tension elements and four bars (compression elements) interconnected by compliant joints. Each bar has a length of L = 11.2 mm, with an initial angle (α) between the bars of 60°. The prestrain in the tension elements is 1.5 mm. The rotational tensegrity joint is a general tensegrity structure, comprising a tension network and two rigid bodies acting as compression elements. As shown in Figure 2b, the length of L_r is 8 mm. Based on the joint equivalence design for tensegrity joints from our previous work,⁵⁴ the rotational tensegrity joint is designed to have a pseudo θ DOF, while the directions of the remaining DOFs are constrained by high stiffness. Additional details regarding the parameters and manufacturing method for the tensegrity joint can be found in Supplementary Section 1.4 in Supplementary Data.

FIG. 2.

Miniature tensegrity joints and bionic tensegrity swimmers. (a) Prismatic tensegrity joint. (b) Rotational tensegrity joint. (c) Schematic diagram and experimental image for magnetic contraction of the prismatic tensegrity joint. (d) Schematic diagram and experimental image for magnetic rotation of the rotational tensegrity joint. (e) Miniature crawling robot. (f) Crawling on the ground. (g) Miniature biomimetic whale with a horizontal tail. (h) Swimming of the miniature biomimetic whale. (i) Miniature biomimetic tuna with a vertical tail. (j) Swimming of the miniature biomimetic tuna.

Both tensegrity joints maintain a self-stressed equilibrium state through their compliant tension network at the neutral pose (zero translations and rotations). The compression elements of both tensegrity joints are fitted with magnetic plates, with their magnetization directions shown in Figure 2a and b. Under the influence of a permanent magnet, the bar angle α of the prismatic tensegrity joint can vary from 60° to approximately 25°, resulting in a contraction of about 40%, as shown in Figure 2c and Supplementary Movie S4. When one end of the rotational tensegrity joint is fixed, it can undergo continuous rotation under the magnetic field, with a maximum rotation range of approximately ±45°, as illustrated in Figure 2d and Supplementary Movie S4.

We next utilize miniature tensegrity joints to construct various miniature tensegrity robots. The detailed parameters of these robots can be found in Supplementary Section 1.5 in Supplementary Data. As shown in Figure 2e, the addition of a “quadruped” configuration transforms the prismatic tensegrity joint into a miniature crawling robot. As shown in Figure 2f and Supplementary Movie S5, by applying a magnetic field, the magnetized crawling robot experiences a magnetic torque, causing its body to actively deform. This results in periodic elongation and contraction, allowing the robot to crawl horizontally on the ground.

Furthermore, by adding a tail, the rotational tensegrity joint can be used to construct two miniature biomimetic swimmers of the similar type. The first one is the miniature biomimetic whale with a horizontal tail, as shown in Figure 2g. Similar to the movement of a whale, by applying a magnetic field, the tail swings up and down, interacting with the fluid to generate propulsion and enabling straight-line swimming, as shown in Figure 2h and Supplementary Movie S5. The second one is the miniature biomimetic tuna with a vertical tail. As shown in Figure 2i and Supplementary Movie S5, by applying a magnetic field, the tail beats horizontally, interacting with the fluid to achieve straight-line swimming. In the future, miniature tensegrity joints could also serve as modular components. By connecting them in series and parallel, more complex miniature tensegrity robots can be constructed to meet a broader range of requirements. The selective and independent control of multiple joints poses a challenge and may be realized through designing the joint’s physical characteristics,⁵⁵ and changing the magnetic field.⁵⁶

Self-adaptive locomotion and excellent environmental adaptability

The miniature tensegrity robot demonstrates remarkable adaptability to various environments and obstacles due to the inherent soft-rigid coupling structure in Supplementary Movies S6 and Movie S7. This capability is demonstrated through the previously developed miniature tensegrity swimmer and spherical tensegrity robot. In addition to the straight-line swimming in the water, by changing magnetic field B, the miniature tensegrity swimmer can execute various planar and 3D motions, including L-shaped diving, double tumbling, autorotation, and revolution, as shown in Figure 3a. As shown in Figure 3b, under single magnetic field control, the compliance of the rotational tensegrity joint enables the miniature swimmer with a length of 14 mm, to swim through obstacles with a gap of 8 mm.

FIG. 3.

Experimental image of miniature tensegrity swimmer. (a) Four different swimming trajectories. (b) Adaptive swimming across underwater terrain. These images have been adjusted to eliminate reflections from the water surface for better clarity.

By altering the rotating magnetic field, the miniature spherical tensegrity robot can also execute S-shaped and ladder-shaped rolling trajectories, as shown in Figure 4a. As shown in Figure 4b, the robot can traverse two different ramps, and perform cross-medium motion, such as land-to-sand and land-to-water. Additionally, as shown in Figure 4c, the robot can navigate various obstacles, including arrayed obstacles (4 mm high, 8 mm spacing), a 45° inclined surface, multilevel steps with 4 mm gaps and 10 mm height, and cylindrical obstacles with a diameter of 10 mm.

FIG. 4.

Experimental images of a miniature spherical tensegrity robot. (a) S-shaped and ladder-shaped rolling trajectories. (b) Adaptive rolling across various terrains, including ramps, and land-to-sand and land-to-water cross-medium transitions. (c) Adaptive rolling through four different obstacles. (d) Adaptive rolling through four types of pipes and channels. (e) Robotic free-fall process from a height of 20 cm.

By applying a simple rotating magnetic field, the miniature spherical tensegrity robot relies on its body compliance to demonstrate efficient navigation in complex environments. Thanks to the passive interaction between the compliant body and the environment, the robot can swiftly navigate through complex, narrow environments (20 mm in width), such as the straight pipe (approximately 7 BL in length), the C-shaped pipe (with a curvature of 0.2 mm⁻¹), the right-angle channel, and an S-shaped channel (with a 20 mm radius curve), as shown in Figure 4d. The time required for these maneuvers is approximately 1.08, 2.03, 1.83, and 1.72 s, respectively. This mechanical intelligence enables the robot to navigate complex environments without requiring collision avoidance algorithms, precise control, active perception, or feedback, thereby reducing the reliance on computational intelligence.

The aforementioned experiment also demonstrates the robot’s collision resistance in pipes and channels, allowing it to survive collisions and recover afterward. The free-fall experiment further highlights the robot’s collision resistance capability. As shown in Supplementary Movie S7, the miniature spherical tensegrity robot undergoes a free fall from heights of 20 cm (about 13 body lengths) and 1.2 m (about 80 body lengths) to the ground. The robot remains structurally intact and undamaged following the collision. Figure 4e illustrates the falling process from a height of 20 cm. Upon impact with the ground, the robot undergoes approximately 50% deformation to absorb the shock, after which it elastically recovers and comes to a stop. These experimental results demonstrate the miniature spherical tensegrity robot’s high impact resistance and energy absorption capabilities.

Tunable stiffness and its effect on swimming velocity

A miniature tensegrity robot can tune its stiffness by changing Young’s modulus and pretension of the tension element.¹⁷ The previously developed miniature whale, constructed by the miniature rotational tensegrity joint, could be employed to demonstrate this ability and function, inspired by the significant influence of the swimmer’s stiffness in the swimming performance.⁵⁷ The miniature swimmer has a fixed body length of 14 mm. The width b of the tension elements in the rotational joint is adjusted (1, 2, and 3 mm) to tune the stiffness of the robotic fish, in Figure 5a. The principle of tunable stiffness is described in Supplementary Section 1.6 in Supplementary Data.

FIG. 5.

Tunable-stiffness miniature tensegrity swimmer. (a) The 3D model and physical prototype of the robot with a marked point. The swimmer’s stiffness is adjusted by changing the width of Elastic-2. (b) Deformation of the miniature swimmer under additional weights. (c) The relationship between the mass of the weights and the rotation angle, which are fitted by the lines in corresponding colors. (d) Swimming velocity of the tensegrity swimmer by varying frequency and stiffness when the magnetic field strength is 80 Gs. (e) Swimming velocity of the tensegrity swimmer by varying magnetic field strength and stiffness at a driving frequency of 20 Hz. (f) Comparison of maximum swimming velocities for tunable-stiffness robotic fish.^{21,32,34,57–68} The swimming experiments of the miniature tensegrity swimmers were repeated at least three times. Error bars represent the standard deviation (SD).

First, the effectiveness of the tunable stiffness is demonstrated. In Figure 5b, the head of the miniature tensegrity swimmer is fixed. The fish body bears its weight (approximately 0.365 g) and varying amounts of additional weights, causing it to rotate. Figure 5c shows the relationship between the rotation angle β of the tensegrity swimmer and the mass of weights, along with the corresponding fitted line. The detailed calculation method for the rotation angle can be found in Supplementary Section 2.4 in Supplementary Data. The fitted line’s slope represents the resistance strength of the miniature swimmer concerning weights, which always indicates the stiffness. The slopes of the fitted lines are denoted as 0.023, 0.019, and 0.010, respectively. As the width b increases, the slope decreases, indicating an increase in stiffness and load capacity. The wide range of slope variation characterizes the effectiveness of the tunable stiffness in both the miniature rotational tensegrity joint and the swimmer.

Previous studies have shown that variable-stiffness characteristics can optimize fluid-structure interaction, improving swimming performance.⁵⁷ We use this characteristic to demonstrate the stiffness variation capability of our miniature tensegrity swimmer. Next, the magnetic field strength B and frequency f were set to 80 Gs and 15–30 Hz, respectively, to investigate the relationship between average swimming velocity and stiffness. The swimming experimental setup and the velocity calculation can be found in Supplementary Section 2.4 in Supplementary Data. In Figure 5d, for the low and medium stiffness swimmers (corresponding to b = 1, 2 mm), the swimming velocity first increases and then decreases with increasing frequency f. For the high-stiffness swimmer (b = 3 mm), the swimming velocity first decreases with increasing f, then increases and gradually declines afterward. The maximum swimming velocity of 3.61 BL/s occurs at high stiffness and f = 20 Hz, while the minimum velocity of 1.46 BL/s is observed at low stiffness and f = 15 Hz. The experimental videos at B = 80 Gs and f = 15 Hz are shown in Supplementary Movie S8. This corresponds to a 147.26% increase in velocity compared with the minimum value. In Figure 5e, the swimming velocity initially increases with magnetic field strength B and then slightly decreases at f = 20 Hz. The high-stiffness swimmer achieves a maximum swimming velocity of 3.67 BL/s, representing a 120.44% increase compared with its minimum value. These nonlinear phenomena reveal the significant impact of tunable stiffness on the swimming velocity.

As shown in Figure 5f and Supplementary Table S1 in Supplementary Data, the maximum swimming velocities of most current tunable-stiffness robotic fish are below 2 BL/s,^{21,32,34–37,59–68} except for a few tunable-stiffness robotic fish. For example, a larval robotic swimmer has a maximum speed of 19.5 BL/s at f = 100 Hz, and a minimum speed of 4.19 BL/s at f = 30 Hz.⁵⁹ Our miniature tensegrity swimmer can achieve a maximum swimming velocity of 3.67 BL/s at f = 20 Hz, which is considered to be at the middle-high level among these tunable-stiffness robotic fish. In addition, most current tunable-stiffness robotic fish’s body length is larger than 26 cm, and they swim in the inertial flow regime (i.e., Reynolds number Re ≥2000).⁵⁹ According to calculations in Supplementary Section 2.3 in Supplementary Data, the miniature tensegrity swimmer’s Reynolds number ranges from approximately 307 to 758 smaller than 2000, indicating that it swims in the intermediate flow regime. In this regime, the viscous forces dominate over the inertial forces, which is quite different from the inertial flow regime. These fluid states exhibit different fluid-structure interactions, resulting in different tunable-stiffness swimming behaviors. In the future, the miniature tensegrity swimmer can help biologists further study the tunable-stiffness swimming in the intermediate flow regime.

Miniature tensegrity robots with high-speed drug delivery and remote heating

The miniature spherical tensegrity robot can accommodate drugs within its internal cavity and may enable drug delivery in the human gastrointestinal tract following future biocompatibility design, as shown in Figure 6a. Experiments tested the drug delivery capability of the robot (see Supplementary Movie S9). The magnetic field strength and frequency were set to 90 Gs and 15 Hz, respectively. Figure 6b shows the relationship between the drug-to-robot mass ratio and its average rolling velocity. Figure 6c shows the relationship between the average rolling speed and magnetic induction intensity of a tensegrity spherical robot with a mass ratio of 0.25 at 15 Hz. The rolling velocity measurement method is similar to that for swimming velocity (see Supplementary Section 2.4 in Supplementary Data). At a mass ratio of 0.25, the robot’s rolling velocity is approximately 25.07 BL/s. The velocity decreases with increasing load. Even with a load 1.25 times its weight, the velocity remains at 14.41 BL/s, demonstrating excellent mobility.

FIG. 6.

Potential medical applications of miniature tensegrity robots. (a) Schematic diagram of gastrointestinal transport. (b) Rolling velocity of the miniature spherical tensegrity robot by varying the mass ratio (transport weight/body weight). The magnetic field strength and frequency were set to 90 Gs and 15 Hz, respectively. Experiments were repeated at least three times. Error bars represent the standard deviation (SD). (c) The relationship between different magnetic field intensity and rolling speed. The magnetic field frequency and mass ratio were set to 15 Hz and 0.25. (d) Maximum velocity of tensegrity robots at different scales.^{18–20,31–45} (e) Relative velocities of some mammals, arthropods, and miniature robots versus body mass. (f) Removal of impurities of different diameters and weights in the pipeline. (g) Temperature sensor. (h) Integration of the thick iron plate in the tensegrity swimmer. (i) Two views of the magnetic induction heating module. (j) Remote heating of the miniature tensegrity swimmer, including a schematic diagram and experimental images.

In Figure 6d and Supplementary Table S2 in Supplementary Data,^{18–20,31–33,38–45,63,69–71} most tensegrity robots are larger than 10 cm in size and have slower velocities (<1.5 BL/s). The rolling velocity of the miniature spherical tensegrity robot is significantly higher than that of other tensegrity robots. Additionally, Figure 6e and Supplementary Table S3 in Supplementary Data present a comparison of relative velocities with respect to body weights including our robots (red stars) and living animals, such as terrestrial arthropods and mammals, as well as previously reported miniature robots.^72–83 The velocities of our robots (represented by five red stars with body masses ranging from 0.424 to 0.954 g, including the weight of the drug) are advanced among miniature robots and mammals, and at a moderate level compared with terrestrial arthropods. The fast locomotion of our robots is advantageous for performing tasks.

The miniature spherical tensegrity robot also exhibits cleaning capabilities. As shown in Figure 6f and Supplementary Movie S9, the robot can propel small spheres made of thermoplastic polyurethane elastomer within the pipeline, to facilitate its cleaning. The spheres’ diameters and weights range from 6 to 16 mm and from 0.108 to 0.932 g (2.20 times the robot weight), respectively. This cleaning capability could be applied in areas such as gastrointestinal foreign body removal and the treatment of mechanical intestinal obstruction. Further miniaturization may enable blood clot removal in vessels.

The soft-rigid hybrid miniature tensegrity robot can utilize a magnetic induction heating module to achieve remote heating in Supplementary Movie S9 while preserving environmental adaptability. Its rigid components can be made of metal, ensuring stable electrical conductivity and geometric properties for increased heating efficiency. The experiment setup is shown in Supplementary Fig. S13 in Supplementary Data. A temperature sensor was made by mixing thermochromic pigment with silicone. It changes from black to red, when heated to above 45°C, as shown in Figure 6g and Supplementary Movie S9. Next, a 0.5-mm thick iron plate was integrated into the rigid part of the previously developed miniature tensegrity swimmer in Figure 6h. The swimmer is placed above the thermochromic temperature sensor and heated by magnetic induction in Figure 6i and j. The contact area between the swimmer and sensor turns red within 30 s, and the heating effect intensifies at 150 s. These experimental results demonstrate that the miniature tensegrity robot has a controllable, noninvasive approach for on-demand localized heating. It shows potential for future clinical applications, such as alleviating gastrointestinal bleeding and providing thermal therapy.

Discussion and Conclusion

In this article, we propose novel soft-rigid hybrid miniature tensegrity robots with a length ranging from 14 to 18 mm, which overcome the size limitations of current tensegrity robots and extend their potential applications. We designed a miniature spherical tensegrity robot inspired by cells and described its assembly and actuation methods. After developing miniature prismatic and rotational tensegrity joints, we constructed both miniature crawling and swimming robots. These examples highlight the wide design and application potential of miniature tensegrity robots.

The miniature tensegrity robot demonstrated high compressibility, collision resistance, tunable stiffness, and strong environmental adaptability through experiments involving static loading, dynamic compression, and high-altitude drop. The robot can withstand dynamic loads of approximately 143,868 times its weight and impact from falls up to 80 body lengths. By leveraging the soft-rigid hybrid nature of the tensegrity structure, the miniature tensegrity robot exhibits multiple modal modes and self-adaptive motion. For example, the miniature tensegrity swimmer features multitrajectory swimming capabilities and can adaptively navigate around obstacles in water. The miniature spherical tensegrity robot can roll across various terrains and obstacles, such as narrow pipes and channels, demonstrating high mobility. The swimming velocity of the miniature tensegrity swimmer ranks middle-high among tunable-stiffness robotic fish. The rolling velocity of the miniature spherical tensegrity robot leads the field of tensegrity robots and miniature robots. These results demonstrate the structural advantages and high locomotion performance of miniature tensegrity robots.

Finally, we explored the potential applications of the miniature tensegrity robot. The miniature tensegrity swimmer swims in the intermediate flow regime. By altering the driving frequency and stiffness, its swimming velocity can be increased by 147.26%. The miniature tensegrity swimmer could help biologists improve their understanding of the mechanisms behind tunable-stiffness swimming in the intermediate flow regime. The miniature spherical tensegrity robot is capable of high-speed drug delivery. Even when carrying a load 1.25 times its weight, the average rolling velocity remains up to 14.41 BL/s. The robot also demonstrates cleaning capabilities, offering the potential for efficient targeted drug delivery and impurity removal in the gastrointestinal tract. Moreover, it can achieve remote heating while maintaining multimodal movement and environmental adaptability. In the future, it may be capable of reaching deeper areas of the human body for remote thermal therapy.

The fabrication method of a miniature tensegrity robot needs to transform the spatial network into a plane topological unit and is not suitable for structures that cannot be transformed. Then it needs to be assembled manually, and it is difficult to assemble at a smaller scale. The existence of closed nodes in the plane topology will cause node errors, and the closed nodes should be minimized to reduce errors. With advanced manufacturing methods (such as two-photon lithography) and bio-compatible design, the size and assemble errors of the miniature tensegrity robot can be further reduced. Due to its high motion performance and multifunctionality, the robot is expected to have wider applications in biomedical and industrial inspection fields. For example, it could be applied in multimedium micro-pipeline inspection and maintenance, microscopic ecosystem or pollutant monitoring, and high-speed targeted treatment of blood vessels and ureters. Biological hybrid robots,⁸⁴ with flexibility, adaptability, biocompatibility, and micro-nano robots integrated with smart materials,⁸⁵ are emerging as new trends in the future of robotic technology.

Footnotes

Authors’ Contributions

B.C. designed and supervised the research. B.C. and Z.H. designed the robots with help from all authors. Z.H., F.Y., Y.Y., and W.H. fabricated the samples and performed the experiments. B.C., Z.H., and F.Y. wrote the article. All authors analyzed the data, discussed the results, and participated in the discussions of the research.

Data Availability

All relevant data supporting the key findings of this study are available within the article and its . Extra data are available from the corresponding author upon reasonable request.

Author Disclosure Statement

The authors declare no competing interests.

Funding Information

This work was supported by the National Natural Science Foundation of China (Grant No. 52205303) and the General Program of the Fujian Provincial Science and Technology Department (Grant Nos. 2023J01056 and 2024J08086).

Supplementary Materials

References

Tang

, et al. A pipeline inspection robot for navigating tubular environments in the sub-centimeter scale. Sci. Robot., 2022; 7:eabm8597.

Mao

, Yang

, Tian

, et al. Magnetic steering continuum robot for transluminal procedures with programmable shape and functionalities. Nat Commun, 2024; 15(1):3759.

Srinivasan

, Dosso

, Huang

H-W

, et al. An ingestible self-propelling device for intestinal reanimation. Sci Robot, 2024; 9(87):eadh8170.

Wang

, Wu

, Yildiz

, et al. Clinical translation of wireless soft robotic medical devices. Nat Rev Bioeng, 2024; 2(6):470–485.

, Yu

, Sun

, et al. Bioinspired claw-engaged and biolubricated swimming microrobots creating active retention in blood vessels. Sci Adv, 2023; 9(18):eadg4501.

Liu

, Xu

, Yang

, et al. Navigation and visual feedback control for magnetically driven helical miniature swimmers. IEEE Trans Ind Inf, 2020; 16(1):477–487.

Ren

, Sitti

. Design and build of small-scale magnetic soft-bodied robots with multimodal locomotion. Nat Protoc, 2024; 19(2):441–486.

, Lum

, Mastrangeli

, et al. Small-scale soft-bodied robot with multimodal locomotion. Nature, 2018; 554(7690):81–85.

Soon

, Yin

, Dogan

, et al. Pangolin-inspired untethered magnetic robot for on-demand biomedical heating applications. Nat Commun, 2023; 14(1):3320.

10.

Skelton

, De Oliveira

. Tensegrity Systems. Springer; 2009.

11.

Kardong

. Vertebrates: Comparative Anatomy, Function, Evolution. McGraw-Hill: New York; 2006; pp. 365–383..

12.

Lyu

, Xu

, Zhou

, et al. Biomimetic thermoresponsive superstructures by colloidal soft-and-hard co-assembly. Sci Adv, 2023; 9(26):eadh2250.

13.

Ingber

. Tensegrity I. Cell structure and hierarchical systems biology. J Cell Sci, 2003; 116(Pt 7):1157–1173.

14.

Liedl

, Högberg

, Tytell

, et al. Self-assembly of three-dimensional prestressed tensegrity structures from DNA. Nat. Nanotechnol., 2010; 5(7):520–524.

15.

Levin

. The tensegrity-truss as a model for spine mechanics: Biotensegrity. J Mech Med Biol, 2002; 02(03n04):375–388.

16.

Scarr

. Simple geometry in complex organisms. J Bodyw Mov Ther, 2010; 14(4):424–444.

17.

Chen

, Jiang

. Instability results from purely rotational stiffness for general tensegrity structure with rigid bodies. Mech. Mach. Theory, 2022; 167:104485.

18.

Böhm

, Kaufhold

, Schale

, et al. Spherical mobile robot based on a tensegrity structure with curved compressed members. In: 2016 IEEE international conference on advanced intelligent mechatronics (AIM). IEEE; 2016; pp. 1509–1514.

19.

Chen

, Jiang

. Swimming performance of a tensegrity robotic fish. Soft Robot, 2019; 6(4):520–531.

20.

, Chen

, Peng

, et al. The equilibrium and form-finding of general tensegrity systems with rigid bodies. Eng. Struct, 2022; 266:114618.

21.

Shintake

, et al. Bio-inspired tensegrity fish robot. In: 2020 IEEE International Conference on Robotics and Automation (ICRA). IEEE; 2020; pp. 2887–2892.

22.

, Vogt

, Rus

, et al. Fluid-driven origami-inspired artificial muscles. Proc Natl Acad Sci U S A, 2017; 114(50):13132–13137.

23.

Shah

, Booth

, Baines

, et al. Tensegrity robotics. Soft Robot., 2022; 9(4):639–656.

24.

Liu

, Bi

, Yue

, et al. A review on tensegrity structures-based robots. Mech. Mach. Theory, 2022; 168:104571.

25.

Mirletz

, Bhandal

, Adams

, et al. Goal-directed cpg-based control for tensegrity spines with many degrees of freedom traversing irregular terrain. Soft Robot, 2015; 2(4):165–176.

26.

Sabelhaus

, et al. System design and locomotion of SUPERball, an untethered tensegrity robot. In: 2015 IEEE international conference on robotics and automation (ICRA). IEEE; 2015; pp. 2867–2873.

27.

Zha

, Wu

, Dimick

, et al. Design and control of a collision-resilient aerial vehicle with an icosahedron tensegrity structure. IEEE/ASME Trans. Mechatron., 2024; 29(5):3449–3460.

28.

Bliss

, Iwasaki

, Bart-Smith

, et al. Central pattern generator control of a tensegrity swimmer. IEEE/ASME Trans. Mechatron., 2013; 18(2):586–597.

29.

Sumi

, Böhm

, Zimmermann

, et al. A multistable tensegrity structure with a gripper application. Mech. Mach. Theory, 2017; 114:204–217.

30.

Friesen

, et al. DuCTT: A tensegrity robot for exploring duct systems. In: 2014 IEEE International Conference on Robotics and Automation (ICRA). IEEE; 2015; pp. 4222–4228.

31.

Liu

, Bi

, Dai

, et al. TICBot: Development of a tensegrity-based in-pipe crawling robot. IEEE Trans. Ind. Electron., 2023; 70(8):8184–8193.

32.

Chen

, Jiang

. Body stiffness variation of a tensegrity robotic fish using antagonistic stiffness in a kinematically singular configuration. IEEE Trans Robot, 2021; 37(5):1712–1727.

33.

Yang

, Yu

, Liu

, et al. A small-scale untethered tensegrity robot with high velocity and multi locomotion modes. IEEE Robot Autom Lett, 2023; 8(11):7503–7510.

34.

Chen

, Wang

, Xiong

, et al. Design and analysis of a novel bionic tensegrity robotic fish with a continuum body. Biomimetics, 2024; 9(1):19.

35.

Tong

, Wu

, Chen

, et al. Design and optimization of an untethered high-performance robotic tuna. IEEE/ASME Trans Mechatron, 2022; 27(5):4132–4142.

36.

Zhong

, Zhu

, Fish

, et al. Tunable stiffness enables fast and efficient swimming in fish-like robots. Sci Robot, 2021; 6(57):eabe4088.

37.

Zou

, Zhou

, Lu

, et al. Tail-stiffness optimization for a flexible robotic fish. Bioinspir Biomim, 2022; 17(6):066003.

38.

Luo

, Wang

, Liu

. Four-bar tensegrity robot based on ADAMS simulation. In: 2017 IEEE International Conference on Mechatronics and Automation (ICMA). IEEE; 2017; pp. 1463–1468.

39.

Böhm

, Zimmermann

. Vibration-driven mobile robots based on single actuated tensegrity structures. In: 2013 IEEE International Conference on Robotics and Automation. IEEE; 2013; pp. 5475–5480.

40.

Cera

, Thompson

, Agogino

. Energy-efficient locomotion strategies and performance benchmarks using point mass tensegrity dynamics. In: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE; 2019; pp. 4678–4683.

41.

Chen

, et al. Inclined surface locomotion strategies for spherical tensegrity robots. In: 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE; 2017; pp. 4976–4981.

42.

Khazanov

, et al. Exploiting dynamical complexity in a physical tensegrity robot to achieve locomotion. Adv Artif Life ECAL, 2013; 12:965–972.

43.

Kim

, et al. Design of a spherical tensegrity robot for dynamic locomotion. In: 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE; 2017; pp. 450–455.

44.

Vespignani

, et al. Design of superball v2, a compliant tensegrity robot for absorbing large impacts. In: 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE; 2018; pp. 2865–2871.

45.

Rieffel

, Mouret

. Adaptive and resilient soft tensegrity robots. Soft Robot, 2018; 5(3):318–329.

46.

Lee

, Elder

, Leal

, et al. A fabrication strategy for millimeter-scale, self-sensing soft-rigid hybrid robots. Nat Commun, 2024; 15(1):8456.

47.

Juan

, Tur

. Tensegrity frameworks: Static analysis review. Mech. Mach. Theory, 2008; 43(7):859–881.

48.

Tur

JMM

, Juan

. Tensegrity frameworks: Dynamic analysis review and open problems. Mech. Mach. Theory, 2009; 44(1):1–18.

49.

Zhang

, Ohsaki

, Tsuura

. Self-equilibrium and super-stability of truncated regular hexahedral and octahedral tensegrity structures. Int. J. Solids Struct, 2019; 161:182–192.

50.

Guest

. The stiffness of prestressed frameworks: A unifying approach. Int. J. Solids Struct, 2006; 43(3–4):842–854.

51.

Chen

, et al. Modular elastic lattice platform for rapid prototyping of tensegrity robots. In: International Design Engineering Technical Conferences and Computers and Information in Engineering Conference. Vol. 58189. American Society of Mechanical Engineers; 2017; p. V05BT08A026.

52.

Lee

, Jang

, Choe

, et al. 3D-printed programmable tensegrity for soft robotics. Sci. Robot., 2020; 5(45):eaay9024.

53.

Ingber

, Wang

, Stamenovic

, et al. Tensegrity, cellular biophysics, and the mechanics of living systems. Rep Prog Phys, 2014; 77(4):046603.

54.

Chen

, Jiang

, Liu

, et al. Joint equivalence design and analysis of a tensegrity joint. J. Mech. Robot, 2021; 13:040904.

55.

Novelino

, Ze

, Wu

, et al. Untethered control of functional origami microrobots with distributed actuation. Proceedings of the National Academy of Sciences of the United States of America, 2020; 117(39):24096–24101.

56.

Wang

, Wu

, Liu

, et al. Selective and independent control of microrobots in a magnetic field: A review. Engineering, 2023; 24:21–38.

57.

Quinn

, Lauder

. Tunable stiffness in fish robotics: Mechanisms and advantages. Bioinspir Biomim, 2021; 17(1):011002.

58.

, Wu

, Nishikawa

, et al. Soft robotic origami crawler. Sci Adv, 2022; 8(13):eabm7834.

59.

Wang

, Ren

, Hu

, et al. Effect of body stiffness distribution on larval fish–like efficient undulatory swimming. Sci Adv, 2021; 7(19):eabf7364.

60.

Chen

, Wu

, Zhang

, et al. Performance improvement of a high-speed swimming robot for fish-like leaping. IEEE Robot. Autom. Lett., 2022; 7(2):1936–1943.

61.

Chen

, Xiong

, Wang

, et al. Performance optimization for bionic robotic dolphin with active variable stiffness control. Biomimetics, 2023; 8(7):545.

62.

, Zhou

, Wang

, et al. Development and stiffness optimization for a flexible-tail robotic fish. IEEE Robot. Autom. Lett., 2022; 7(2):834–841.

63.

, Zhou

, Wang

, et al. Toward swimming speed optimization of a multi-flexible robotic fish with low cost of transport. IEEE Trans Automat Sci Eng, 2024; 21(3):2804–2815.

64.

, Jiang

, Wang

, et al. A soft robotic fish with variable-stiffness decoupled mechanisms. J Bionic Eng, 2018; 15(4):599–609.

65.

Liao

, Zhou

, Wang

, et al. A wire-driven dual elastic fishtail with energy storing and passive flexibility. IEEE/ASME Trans Mechatron, 2024; 29(3):1914–1925.

66.

Liao

, et al. A fast online elastic-spine-based stiffness adjusting mechanism for fishlike swimming. Soft Robot, 2024; doi: 10.1089/soro.2023.0204

67.

Liu

, Jiang

, Xu

. Development of novel fish-inspired robot with variable stiffness. Ocean Eng, 2024; 305:118047.

68.

Qiu

, Wu

, Wang

, et al. Locomotion optimization of a tendon-driven robotic fish with variable passive tail fin. IEEE Trans. Ind. Electron., 2023; 70(5):4983–4992.

69.

Rovira

, Tur

JMM

. Control and simulation of a tensegrity-based mobile robot. Robot. Auton. Syst, 2009; 57:526–535.

70.

Wang

, Ling

, Qiu

, et al. A four-prism tensegrity robot using a rolling gait for locomotion. Mech. Mach. Theory, 2022; 172:104828.

71.

Wang

, Li

, He

, et al. A light‐powered ultralight tensegrity robot with high deformability and load capacity. Adv Mater, 2019; 31(7):e1806849.

72.

Rubin

, et al. Exceptional running and turning performance in a mite. J. Exp. Biol, 2016; 219:676–685.

73.

Duncan

, et al. A comparison of the energetics of foraging of three species of Leptogenys (Hymenoptera, Formicidae). Physiol. Entomol, 1993; 18:372–378.

74.

Alexander

, Langman

, Jayes

, et al. Fast locomotion of some African ungulates. J. Zool, 1977; 183(3):291–300.

75.

Chen

, Yang

, Zhang

, et al. Scale-inspired programmable robotic structures with concurrent shape morphing and stiffness variation. Sci Robot, 2024; 9(92):eadl0307.

76.

Chi

, Hong

, Zhao

, et al. Snapping for high-speed and high-efficient butterfly stroke–like soft swimmer. Sci Adv, 2022; 8(46):eadd3788.

77.

, Liu

, Cacucciolo

, et al. An autonomous untethered fast soft robotic insect driven by low-voltage dielectric elastomer actuators. Sci Robot, 2019; 4(37):eaaz6451.

78.

, Kim

, et al. High-performance electrified hydrogel actuators based on wrinkled nanomembrane electrodes for untethered insect-scale soft aquabots. Sci Robot, 2022; 7(71):eabo6463.

79.

Liang

, Wu

, Yim

, et al. Electrostatic footpads enable agile insect-scale soft robots with trajectory control. Sci Robot, 2021; 6(55):eabe7906.

80.

Liu

, Feng

, Cheng

, et al. Singularity analysis and solutions for the origami transmission mechanism of fast-moving untethered insect-scale robot. IEEE Trans Robot, 2024; 40:777–796.

81.

Liu

, Zhan

, Liu

, et al. A wireless controlled robotic insect with ultrafast untethered running speeds. Nat Commun, 2024; 15(1):3815.

82.

Ramezani

, Chung

S-J

, Hutchinson

, et al. A biomimetic robotic platform to study flight specializations of bats. Sci. Robot., 2017; 2(3):eaal2505.

83.

, Yim

, Liang

, et al. Insect-scale fast moving and ultrarobust soft robot. Sci Robot, 2019; 4(32):eaax1594.

84.

Mestre

, Patiño

, Sánchez

. Biohybrid robotics: From the nanoscale to the macroscale. Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology, 2021; 13:e01703.

85.

Soto

, Karshalev

, Zhang

, et al. Smart materials for microrobots. Chem Rev, 2022; 122(5):5365–5403.