There is strong motivation for tactile display technology capable of producing softness and thermal sensations simultaneously. Such a device could simulate thermal properties of vascular disease in medical training, or the varying stiffness and thermal emissivity of different materials in a virtual reality simulation. This work presents a multimodal surface that integrates constrictive particle jamming with a fluidic thermal skin. Its contributions are a demonstration of how pump-free jamming can mediate otherwise incompatible technologies, a mechanical design that optimizes thermal transfer without compromising jamming compliance, and an experimental characterization of the device’s performance.
Temperature Interface
Thermal feedback is provided by a thin silicone skin attached to the top surface of the jamming container. Channels that route liquid through the skin are cast from Smooth-On Dragon Skin 10 FAST and attached to a touch surface made from a composite of Dragon Skin 10 FAST and copper powder, with a mesh size of 325 μm, at a 60:40 ratio by mass. The copper composite layer provides a thermally conductive interface between adjacent channels, the temperature-controlled liquid, and a user’s finger. The pure silicone insulates the large thermal mass of the particle fluid.
Performance and Characterization
The device’s performance was characterized in terms of the range of achievable softness change and the range and consistency of its surface temperature.
Softness
The stiffness of the display was analyzed as a function of motor torque, using closed-loop current control. Torque was varied from 0 to 2 Nm in increments of 0.2 Nm.
Temperature
Closed-loop performance was assessed by commanding the display to reach and hold a moderate setpoint temperature. The time taken to reach the temperature was measured, as was the peak error once the setpoint was reached. All measurements were taken with the system described above.
Thermal Response
To evaluate closed-loop performance, the display was heated to approximately 33°C and then given a setpoint of 15°C. It took 12 minutes to reach the setpoint, an average temperature change of 1.5°C per minute. After reaching the setpoint, the measured temperature remained within ±0.2°C of it for 15 minutes.
Throughout all tests, a 30-second latency was observed between sending a command to the display and observing the expected heating or cooling response.
Discussion
The prototype demonstrated a good range of achievable stiffness, despite initial contact stiffness being more susceptible to the low surface stiffness of the thermal skin. Stiffness was also highly controllable, with consistent responses to torque in the 0–1 Nm range. Hysteresis was initially high because the thermal skin’s elasticity could build potential energy when depressed into the softer jamming substrate, but it decreased quickly and was consistent thereafter. This would have a noticeable effect on tactile sensation.
Surface uniformity was generally very good, with little evidence of the water losing energy as it flowed through the thermal skin. Temperature was also generally consistent above and between channels, indicating that the copper-silicone composite effectively spread thermal energy. External thermal photographs indicated that the insulation between the thermal layer and the jamming substrate was effective, suggesting that the material design could concentrate heat transfer at the display.
Preliminary evaluation of closed-loop control demonstrated sub-degree deviation, typically ±0.2°C. More in-depth testing is required to validate this performance robustly. Recovery from perturbation was not tested. The dynamic response was somewhat weak given the 30-second latency in temperature change. This is likely due to poor thermal inertia in the Peltier module and fluid system, which itself has a relatively high thermal mass.
Applications requiring faster hot-cold changes may benefit from dedicated hot and cold water supplies, with either hot or cold water pumped depending on the target temperature. The system as presented is usable in perceptual research and medical simulation, which typically would not require rapid temperature changes.
Future Work
Future work will investigate the benefits of smaller display areas to enhance the response rate. Dynamic response will be improved by reducing the display size and enhancing liquid temperature control.
Scientific work will apply the interface to address known challenges in creating multisensory experiences and investigate crossmodal correspondences involving both softness and temperature. These experiences include medical simulation, where softness and temperature (heat or cold) are important indicators of infection, malignancy, and vascular disease, as well as affective haptic experiences that benefit from softness and warmth.
Further investigations will explore crossmodal perception in haptic interactions, including recognition of physical materials by their compliance and thermal conductivity, and interactions among softness, temperature, and weight—a variation of Weber’s Illusion.
Conclusion
This work demonstrates that particle jamming can serve as a foundation for multimodal tactile interfaces that combine otherwise incompatible technologies to produce multiple distinct sensations simultaneously. This is demonstrated by combining softness and temperature change in a desktop-scale haptic device. Future research will aim to improve dynamic performance and use this capability to explore the perceptual characteristics and application areas of these crossmodal haptic interactions.
This work was supported by the Imperial College-MIT Global Seed Fund and EPSRC pro2 Network+, grant number EP/W020564/1. J. Brown was supported by the Imperial College Digital Innovation Fund.
Frequently Asked Questions
How does the device produce softness and temperature sensations at the same time? It combines constrictive particle jamming with a fluidic thermal skin attached to the jamming container.
What material helps spread heat across the touch surface? A copper-silicone composite layer provides a thermally conductive interface between the fluid channels and the user’s finger.
How quickly can the display change temperature? In the reported test, it took 12 minutes to cool from approximately 33°C to a 15°C setpoint, with a 30-second latency observed between commands and the expected response.
What applications could use this multimodal interface? The paper identifies medical simulation and perceptual research, including experiences involving vascular disease, material recognition, and affective haptics.