A Simple Manufactured Hardness Sensor for Surgical Robotics

Robotic surgery has advanced fast in the last decade, but one thing it still lacks is a true sense of touch and surgeons still depend on tactile cues every time they handle tissue. Robots, on the other hand, can cut and move with high precision, however, they have no way to “feel” what they touch. That absence changes the whole character of an operation. To develop sensors that can mimic this sense has turned out to be much harder than it first appeared because biological tissue isn’t a simple material; it reacts, relaxes, and changes shape depending on how force is applied. Skin and organs are both viscoelastic, which means their mechanical behavior depends on time as well as stress. Therefore, a single measurement rarely tells the full story. Traditional tests—nanoindentation, elastography, aspiration—can give numbers for stiffness or elasticity, but they require bulky instruments and controlled conditions. None of that works well inside a surgical robot or on living tissue that’s moving and wet. Even the latest tactile sensors which are considered sensitive, often mix up deformation and pressure signals. They can register that something happened, but not what kind of force caused it. It is still getting a robot to sense hardness the way a human finger does remains one of the biggest unsolved problems in the field.

To this account, new research paper published in Advanced Electronic Materials and conducted by Associate Professor Tao Yue, Dr. Yuyin Zhang, Dr. Yuanjie Gan, Dr. Chengzhi Hu, and Assistant Professor Yue Wang from the Shanghai University, the researchers developed a multi-layer liquid metal-based composite sensor capable of simultaneously measuring pressure and strain in a fully decoupled manner. Two integrated resistive sensing models—a micro-pump pressure sensor and a grid-structured strain sensor—were fabricated via one-step continuous injection of gallium–indium alloy into laminated PDMS channels. The research team employed a continuous liquid metal injection process to construct a multilayer tactile sensor composed of laminated PDMS microfluidic layers. Each layer was designed to perform a distinct sensing function: one detecting normal pressure, the other measuring surface strain. This decoupled arrangement allowed the sensor to interpret hardness through the ratio of applied force to induced deformation. The top layer acted as a pressure sensor incorporating a micro-pump-like channel network, while the lower layer adopted a Wheatstone bridge configuration to register strain. Both sensors relied on liquid gallium–indium alloy as the conductive element, chosen for its stability, low viscosity, and high conductivity within soft polymer matrices.
During fabrication, the authors used two syringes to inject the liquid metal simultaneously through microchannels to prevent bubble entrapment and ensure continuous conductivity across the structure. Plasma bonding enabled tight lamination between PDMS layers, resulting in a compact 25 mm device with high mechanical stability. Unlike multi-material composites that often suffer from delamination under load, this integrated method produced a monolithic structure capable of sustaining repeated deformations without performance degradation. They also conducted finite-element simulations which confirmed the deformation patterns of both micro-pump and strain-grid layers, and showed how localized stress altered channel geometry and resistance. They tested experimentally and found the pressure and strain sensors exhibited clear functional independence. Under static loading with 100 g weights, the pressure sensor displayed a prompt and repeatable resistance change upon application and release, with response times of 589 ms and 359 ms, respectively. In contrast, the strain sensor showed minimal fluctuation under similar loading but responded linearly to tensile strain between 4% and 12%, confirming effective decoupling. Fatigue tests involving 1000 cycles of pressure or strain showed no statistically significant drift in signal, validating the device’s mechanical endurance. They also applied the sensor to various materials—glass, PDMS, and silicone—to emulate biological hardness differences. Results indicated that harder surfaces produced smaller strain-to-pressure ratios, whereas softer materials, akin to muscle or skin, showed greater deformation responses. Integration onto robotic and human fingertips revealed real-time tactile mapping capabilities, with distinct signal clusters for different surfaces. Furthermore, the team miniaturized the system into a 2×2 sensor array, illustrating potential for spatial mapping and the capture of fine surface morphology. A dedicated PCB incorporating STM32 microcontrollers and OLED output enabled direct digital readout, confirming the feasibility of closed-loop tactile feedback for surgical robotics.

In conclusion, the study by Shanghai University scientists successfully developed a new simple and powerful framework for tactile sensing in medical robotics. the researchers addressed one of the most persistent challenges in tactile sensor design—achieving true decoupling between normal and tangential mechanical stimuli by merging resistive pressure and strain sensing into a single laminated microfluidic architecture. This innovation enables reliable detection of material hardness using a device no thicker than a fingertip layer, fabricated through accessible, low-cost soft lithography techniques. The simplicity of the design stands in contrast to prior systems that rely on complex electrode deposition, multiple composites, or fragile capacitive elements. Beyond fabrication convenience, the scientific importance of this work lies in its direct applicability to surgical robotics, where distinguishing between tissues of subtly different stiffness can influence intraoperative safety and outcomes. A robot equipped with such a sensor could, in principle, differentiate between healthy and pathological tissues, guide delicate resections, or provide quantitative feedback during minimally invasive procedures. In addition, the thin, flexible format makes it compatible with prosthetic hands, rehabilitation devices, and wearable healthcare monitors. The researchers’ demonstration of an array configuration further points toward tactile imaging—essentially enabling robotic fingers to “feel” texture and compliance distributions across surfaces. From a materials perspective, the use of gallium–indium liquid metal represents a particularly elegant choice. Its self-healing electrical continuity and mechanical resilience allow the sensor to maintain stable output under repeated deformation, a limitation that often plagues solid metallic or piezoresistive counterparts. Moreover, the continuous injection fabrication ensures uniformity and scalability, suggesting a path toward batch production or even integration with stretchable circuit boards. Looking ahead, the concept could evolve into high-resolution tactile skins capable of mapping three-dimensional compliance landscapes or identifying hidden anomalies beneath soft surfaces. In medical contexts, such technology might eventually contribute to non-invasive diagnostic procedures, such as detecting fibrotic or tumorous tissue regions through mechanical contrast. More broadly, the new study exemplifies how simplicity in design when grounded in sound mechanical and material reasoning can produce highly functional systems and move surgical robotics closer to the dexterity and perception of the human hand.