Bendable Bots: 3D-Printed Structures Enable Highly Sensitive Soft Robotics

Author: Denis Avetisyan

Researchers have developed a new method for creating sensorized soft robotic segments using 3D printing and embedded fluidic channels, paving the way for more adaptable and responsive robots.

This soft continuum robot utilizes helicoid segments with integrated air channels to achieve multi-material functionality and adaptable movement.

This work details the fabrication and control of meter-scale soft continuum robots with integrated deformation and tactile sensing through multi-material 3D printing of helicoid structures.

While soft robots offer adaptability and safety for complex environments, their inherent deformability poses challenges for precise sensing and control. This is addressed in ‘Printed helicoids with embedded air channels make sensorized segments for soft continuum robots’, which introduces a novel fabrication method integrating fluidic innervation and 3D printing to create sensorized soft robotic segments. The approach utilizes helicoid lattices with embedded air channels and miniature sensors, demonstrated in a meter-scale, 14-DoF arm capable of trajectory tracking, object grasping, and tactile-based stiffness detection. Could this scalable fabrication strategy unlock new possibilities for fully-integrated, high-performance soft robotic systems in diverse applications?

Beyond Rigidity: Embracing Compliant Sensing

Conventional robotics, built upon frameworks of rigid materials and reliant on a limited number of discrete sensors, often struggles when confronted with the unpredictable nature of real-world environments. This design philosophy necessitates precise modeling and control, yet fails to account for the inherent uncertainties of interaction, leading to potential damage to both the robot and its surroundings. The inflexibility of these systems hinders their ability to adapt to unforeseen obstacles, delicate manipulations, or dynamic changes in the workspace, ultimately limiting their application in scenarios demanding close collaboration with humans or operation within unstructured settings. Consequently, advancements are crucial to move beyond systems that simply react to contact and towards those capable of proactively sensing and responding to the complexities of a dynamic environment.

Existing sensing technologies, while well-established in traditional robotics, present significant challenges when applied to the emerging field of soft robotics. Optical sensors, susceptible to occlusion and requiring clear lines of sight, struggle with the inherent deformability of soft materials. Capacitive and resistive sensors, though promising, often exhibit limited sensitivity and are prone to drift, particularly under the continuous strain and complex deformations characteristic of soft robots. Furthermore, integrating these discrete sensors into the very fabric of a soft robot-essential for truly distributed sensing-proves difficult without compromising the material’s flexibility and responsiveness. These limitations collectively hinder the development of soft robots capable of nuanced interaction and reliable operation in unstructured environments, necessitating the exploration of novel sensing paradigms.

The future of robotics hinges on a shift towards systems capable of sophisticated and safe collaboration with humans, necessitating a fundamental change in how robots perceive their surroundings. Traditional robots, built on rigid frames and limited sensing, struggle to navigate unpredictable environments and interact delicately with people or fragile objects. Instead, researchers are exploring robots with intrinsic sensing – the ability to ‘feel’ through their own deformable bodies – and distributed sensing, where perception isn’t limited to discrete points but is woven throughout the robot’s structure. This approach allows for a more nuanced understanding of contact, force, and texture, moving beyond simple collision detection to enable robots to adapt to complex interactions and operate safely in close proximity to humans, mirroring the sensitivity and dexterity found in biological systems.

Robotics is evolving beyond merely identifying that contact has occurred, and increasingly focuses on discerning how objects interact through deformation. This shift necessitates sensors capable of resolving not just pressure, but also shear forces, curvature, and strain distribution across a robot’s surface. Researchers are developing materials and algorithms that interpret subtle changes in shape as nuanced information about an object’s weight, texture, or even internal structure. This ability to ‘read’ deformation allows for more delicate manipulation, improved grasp stability, and safer human-robot collaboration, as the robot can dynamically adjust its actions based on a comprehensive understanding of the physical interaction. Ultimately, this move towards deformation-based sensing promises a new generation of robots capable of truly adaptive and intelligent behavior.

Average air pressure readings from the robot's contact sensors reliably differentiate objects based on their stiffness, as indicated by the standard deviation shown in the error bars. — Average air pressure readings from the robot’s contact sensors reliably differentiate objects based on their stiffness, as indicated by the standard deviation shown in the error bars.

Fluidic Innervation: Sensing Through Embodied Intelligence

Fluidic innervation utilizes a network of microfluidic channels integrated directly into the body of a soft robot. These channels are filled with an incompressible fluid, and deformation of the robot’s structure causes corresponding changes in the channel volume and, consequently, internal fluid pressure. By strategically embedding these channels and precisely measuring the resulting pressure variations at multiple points, the system can accurately map the 3-dimensional deformation of the robot. This approach enables the robot to sense external forces and contact without relying on external sensors, offering a fully embedded and scalable sensing solution.

Pressure variations within the fluidic channels of a fluidic innervation system directly correlate to external forces applied to the soft robot’s structure. As the robot deforms due to contact or interaction with its environment, the volume of the embedded fluid channels changes. These volumetric changes are registered as measurable pressure differentials by integrated pressure sensors. The magnitude and location of these pressure changes provide data regarding the intensity and spatial distribution of the applied forces, effectively allowing the robot to ‘feel’ contact, determine the magnitude of applied loads, and perceive its interaction with surrounding objects. This pressure data is then processed to generate a representation of the external forces acting upon the robot.

Fluidic innervation utilizes the compressibility of liquids to achieve highly sensitive and distributed deformation sensing within soft robotic systems. Microfluidic technologies, specifically the fabrication of miniature channels and reservoirs, enable the integration of these fluid-filled networks directly into the robot’s body. This scalability allows for a high density of sensing points without significantly increasing the robot’s weight or complexity. Because fluids readily conform to changes in shape, pressure variations within the embedded channels directly correlate to local strain and deformation, providing a detailed map of the robot’s physical interaction with its environment. The inherent compliance of the fluid medium also enhances the system’s robustness against noise and minor structural imperfections.

Inertial Measurement Units (IMUs) are integrated with fluidic innervation systems to provide complementary kinematic data essential for accurate state estimation. While fluidic innervation detects local deformation and provides feedback on external forces, it lacks absolute positional or rotational awareness. IMUs, consisting of accelerometers and gyroscopes, directly measure linear acceleration and angular velocity, respectively. Fusing this data with the pressure readings from the fluidic channels allows for a more complete understanding of the robot’s pose and movement in three-dimensional space. This integration is particularly important for dynamic applications, enabling closed-loop control strategies that account for both deformation and overall robot motion, improving stability and responsiveness.

Characterization of the embedded air channel sensors reveals pressure and IMU responses that vary predictably with applied loading conditions.

Precision Fabrication: Architecting Softness with Vision-Guided Jetting

Vision-Controlled Jetting (VCJ) is an additive manufacturing process capable of depositing multiple materials with sub-millimeter precision. This technique facilitates the creation of complex geometries, specifically enabling the fabrication of soft robotic components with integrated fluidic channels. VCJ utilizes a camera-based feedback system to monitor material deposition in real-time, correcting for deviations and ensuring accurate placement of both compliant and rigid materials. This capability allows for the direct 3D printing of sealed, internal channels within soft robot structures, eliminating the need for post-fabrication assembly or bonding of separate channel components and reducing potential failure points.

The fabrication process leverages the distinct mechanical properties of Thiol-ene Polyurethane Elastomer (TEPU) and Epoxy resins to achieve functional heterogeneity in soft robotic components. TEPU, a soft, compliant material, is utilized for sections requiring flexibility and deformation, enabling movement and adaptability. Conversely, Epoxy, a rigid thermoset polymer, is employed for structural elements demanding high strength and dimensional stability. This dual-material approach allows for the creation of soft robots with integrated rigidity, facilitating the construction of complex geometries and the precise placement of functional components like sensors and fluidic channels within a single fabrication run.

Precise material deposition is achieved through a vision-controlled jetting process, enabling the fabrication of complex internal channel networks within soft robotic structures. This technique allows for the selective placement of fluids – specifically, TEPU and epoxy – with micron-level resolution, facilitating the creation of interconnected pathways for pneumatic or hydraulic actuation. The system utilizes real-time visual feedback to monitor deposition accuracy and adjust jetting parameters, ensuring channel dimensions and configurations adhere to the design specifications. This level of control is critical for embedding functional elements, such as sensors and actuators, directly into the robot’s body and creating the necessary fluidic logic for sophisticated movements and sensing capabilities.

A 1-meter long, 14-degrees-of-freedom (DoF) soft robot arm was successfully fabricated using vision-controlled jetting techniques. This arm incorporates integrated sensing capabilities directly within its structure during the fabrication process. The fabrication demonstrates the scalability of the process to produce large-format soft robots with complex articulation and the ability to embed functional elements, such as sensors, without post-assembly procedures. The resulting arm’s dimensions and DoF count represent a significant advancement in the complexity and functionality of fabricated soft robotic systems.

The N6 segment features a helicoid structure with integrated air channels and rigid end plates to facilitate airflow and structural support.

Towards Adaptive Manipulation: Embodying Intelligent Touch

A novel robotic manipulator combines the precision of cable-driven mechanisms with the adaptability of fluidic innervation, resulting in a system capable of nuanced interaction with diverse objects. This architecture utilizes soft, fluid-filled chambers as artificial muscles, controlled by pressurized air, to achieve a high degree of flexibility and compliance. Unlike traditional rigid robots, this design allows the manipulator to conform to the shape of grasped objects, distributing forces evenly and preventing damage. The cable-driven framework provides precise control over the fluidic actuators, enabling complex movements and delicate manipulation tasks. This integration effectively bridges the gap between the strength and accuracy of conventional robotics and the adaptability found in biological systems, paving the way for more robust and versatile robotic solutions.

The robot achieves delicate manipulation of fragile objects through an integrated pressure sensing system that effectively discerns object stiffness. By monitoring the distribution of force applied during grasping, the robot can dynamically adjust its grip strength and conform to the object’s shape, preventing damage. This tactile feedback allows it to differentiate between objects requiring a gentle touch-such as a ripe fruit or a delicate flower-and those that can withstand a firmer hold. The system doesn’t rely on pre-programmed assumptions about an object’s fragility; instead, it feels its way to a secure grip, ensuring both stability and safety during manipulation. This capability moves beyond simple pick-and-place tasks, enabling the robot to interact with a wider range of objects in unstructured environments and perform tasks requiring a nuanced understanding of material properties.

The robot achieves remarkably precise and controlled movements through a sophisticated integration of deformation feedback and trajectory tracking. Rather than relying solely on pre-programmed motions, the system continuously monitors how the gripper deforms as it interacts with an object. This real-time data-measuring the subtle bending and flexing of the soft robotic components-is then fed back into the trajectory tracking algorithm. The algorithm dynamically adjusts the robot’s movements, compensating for unexpected resistance or variations in object shape. This closed-loop control system enables the robot to not only follow a desired path but also to maintain a consistent grip force and avoid damaging delicate objects, effectively ‘feeling’ its way through complex manipulation tasks with a level of finesse previously unattainable in conventional robotic systems.

The robotic system exhibits a capacity for tactile discernment, effectively differentiating between soft and rigid objects through analysis of pneumatic pressure data. By monitoring the average pressure readings across multiple air channels embedded within the gripper, the system can reliably categorize objects based on their resistance to deformation. This nuanced feedback allows for the implementation of adaptive grasping strategies; the robot automatically adjusts its grip force and manipulation approach based on the perceived stiffness. Consequently, the system demonstrates the potential to handle a diverse range of objects, from delicate produce to robust tools, without causing damage or compromising stability – a crucial step towards more versatile and intelligent robotic manipulation.

A custom printed circuit board integrating pressure sensors and an inertial measurement unit is paired with a CAN-USB adapter to facilitate data transfer to a host computer.

Architected Compliance: The Future of Robot Interaction

The emerging field of soft robotics is being revolutionized by the synergy between fluidic innervation and architected materials, notably structures like the Helicoid. This innovative combination allows for a degree of control over a robot’s compliance and deformation previously unattainable with rigid systems. By embedding fluidic channels within specifically designed, geometrically complex soft materials, engineers can precisely dictate how a robot bends, twists, and responds to external forces. Unlike traditional actuators, this approach doesn’t rely on bulky motors and gears, instead leveraging the principles of fluid mechanics to achieve nuanced, adaptable movements. The result is a new generation of robots capable of navigating unstructured environments, delicately manipulating objects, and, crucially, interacting with humans in a safer and more intuitive manner, as subtle changes in fluid pressure translate to predictable and controllable physical responses.

Robots built with fluidic innervation and architected soft materials demonstrate a capacity for safe and effective navigation of challenging environments. These designs prioritize adaptability through controlled deformation, allowing robots to absorb impacts and maneuver around obstacles, unlike rigid systems. This inherent compliance is particularly crucial for human-robot interaction, minimizing the risk of injury during collaborative tasks or accidental contact. By mimicking the flexibility of natural systems, these robots can operate in spaces and alongside people with increased safety and efficiency, opening possibilities for applications in healthcare, search and rescue, and even everyday assistance within domestic settings.

Rigorous testing has demonstrated a high degree of accuracy in predicting the mechanical behavior of novel, helicoidally-structured soft robots. Specifically, computational models accurately forecasted axial and bending stiffness across four distinct designs – N4, N4T, N6, and N8 – with measured values falling within a tight ±15% margin of error. This close correlation between prediction and physical performance confirms the efficacy of the design and fabrication methodology, paving the way for reliable and repeatable creation of compliant robotic structures. The validation suggests that future iterations and more complex designs can be confidently modeled and prototyped, accelerating the development of robots capable of nuanced interaction and adaptability in real-world environments.

The trajectory of robotics is shifting from a focus on sheer power and precision to one prioritizing adaptability, safety, and intuitive interaction. This emerging paradigm envisions robots seamlessly integrating into human environments, not as rigid automatons, but as compliant and intelligent collaborators. Such machines will leverage advancements in materials science and artificial intelligence to dynamically adjust to unforeseen circumstances, navigate complex terrains, and respond to human needs with nuanced understanding. Ultimately, the goal is to create robotic systems that are not simply capable of performing tasks, but are also trustworthy partners, fostering a future where humans and robots coexist and collaborate effectively.

Experimental measurements of axial and bending stiffness for four helicoid designs (N4, N4T, N6, N8) closely align with theoretical predictions, validating the design methodology.

The pursuit of increasingly complex robotic systems often obscures a fundamental principle: effective design stems from minimizing unnecessary elements. This research, detailing the fabrication of sensorized soft continuum robots through multi-material 3D printing, exemplifies this ethos. The integration of fluidic innervation and helicoid structures, while innovative, serves a clear purpose – providing deformation feedback and tactile sensing – without extraneous complication. As Robert Tarjan observed, “Simplicity is prerequisite for reliability.” The presented methodology achieves a functional, meter-scale robot not through architectural extravagance, but through a focused application of materials and design, prioritizing clarity in both structure and function. The core idea centers around the reduction of complexity in achieving robust sensor integration.

Further Refinements

The demonstrated fabrication-architected helicoids with fluidic channels-solves a specific engineering problem. However, the underlying principle reveals a broader, and perhaps more troubling, truth: complexity is often introduced not by necessity, but by the limitations of available manufacturing processes. The current iteration relies on multi-material printing, a technique inherently constrained by resolution and material compatibility. Future work should address the possibility of functionally equivalent structures achievable through simpler fabrication methods – a reduction, not an addition, of complexity.

The meter-scale demonstration is merely a scaling exercise. The true challenge lies not in size, but in robustness and longevity. Degradation of the elastomeric materials, clogging of the fluidic channels, and the inevitable failure of sensor elements represent critical limitations. The pursuit of more exotic materials-self-healing polymers, perhaps-is a distraction. The solution likely resides in a re-evaluation of the sensor architecture itself, favoring passive, mechanically robust designs over active, electronically dependent systems.

Ultimately, the value of this work is not in the creation of a particular robot arm, but in the articulation of a design philosophy. Emotion is a side effect of structure; a successful soft robot will not mimic biological systems, but will exceed them through elegant simplicity. Clarity is compassion for cognition. The field must prioritize fundamental understanding over superficial imitation.

Original article: https://arxiv.org/pdf/2602.23457.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/