Tsinghua University Works Out How to Sort the Rough from the Smooth, with Robotics

Researchers at Tsinghua University developed a robotic sensor mimicking human touch, enhancing remote scientific experiments by accurately identifying materials through thermal and tactile sensing with an impressive accuracy.

Researchers at Tsinghua University have developed an innovative tactile sensing technology that could significantly enhance the capabilities of robots used in remote scientific experiments, according to a recent study published in Applied Physics Reviews by AIP Publishing.

The new technology mimics human touch sensing, particularly the ability to perceive thermal changes, enabling robots to distinguish between objects that are similar in shape and size but differ in material properties. This breakthrough addresses previous limitations where robotic sensors often struggled with objects that were visually or texturally similar. We’ve previously tried out some commercial examples which while impressive were still texturally limited.

Rong Zhu, the lead author of the study, explains, “utilizing spatiotemporal tactile sensing during hand grasping to extend the robotic function and ability to simultaneously perceive multi-attributes of the grasped object, including thermal conductivity, thermal diffusivity, surface roughness, contact pressure, and temperature”.

The researchers created a layered sensor system that is revolutionary in its approach to object recognition. The top layer of the sensor is designed for material detection, while the bottom layer measures pressure. Between these layers, a porous material reacts to thermal changes, enhancing the sensor’s ability to accurately identify various objects.

To demonstrate the effectiveness of this new sensor, the Tsinghua team equipped a robot with their technology to sort rubbish. The robot successfully identified and sorted a wide range of rubbish items — from orange peels to plastic bottles — with a remarkable accuracy rate of 98.85%. This capability not only suggests applications in waste management, but also in complex remote scientific experiments where precise material identification is crucial.

For instance, in remote environmental monitoring or space exploration, robots could utilize this technology to analyse soil samples or collect and categorize extraterrestrial materials. The ability to accurately identify these materials remotely could reduce the need for physical sample returns, speeding up data collection and reducing costs.

The potential for remote application of this technology in scientific experiments is immense. By allowing for more precise control and identification capabilities, researchers can conduct more detailed and varied experiments from afar. This not only enhances the efficiency of data collection but also opens up new possibilities for experiments in environments that are inhospitable or inaccessible to humans.

As the technology develops, future research will focus on refining these robotic systems to enhance their autonomy and embodied intelligence, potentially transforming both how we interact with robots and their role in scientific discovery.

You can read the full paper “Enhanced robotic tactile perception with spatiotemporal sensing and logical reasoning for robust object recognition” here in Applied Physics Reviews.

Staff Writer

Our in-house science writing team has prepared this content specifically for Lab Horizons

Leave a Reply

Your email address will not be published. Required fields are marked *