Replicating the human somatosensory network in robots is crucial for dexterous manipulation to ensure appropriate grasping forces for objects with different softness and textures. Although artificial tactile perception has made progress in object recognition, accurately quantifying tactile perception for softness and texture identification remains challenging.
To address this problem, recently, the team of Wu Huaping from Zhejiang University of Technology and Jiang Hanqing from Westlake University reported a method of using a dual - mode tactile sensor to capture multi - dimensional static and dynamic stimuli, allowing for the simultaneous quantification of softness and texture characteristics. This method enables the synergistic measurement of elastic and frictional coefficients, thus providing a general strategy for obtaining the adaptive gripping force necessary for scar - free and anti - slip interactions with delicate objects. By equipping this sensor, the robotic arm can identify porcine mucosal features with an accuracy of 98.44% and stably pick ripe white strawberries that are visually indistinguishable, enabling reliable tissue palpation and intelligent harvesting. The proposed design concept and comprehensive guidelines will provide insights for the development of tactile sensors and bring prospects for robotics.
Figure 1. Dual - mode tactile sensor for softness and texture identification inspired by the human fingertip
Figure 2. Structure, sensing principle and multi - dimensional force - sensing performance of the dual - mode tactile sensor
Figure 3. Softness identification of the tactile sensor based on dynamic classification and static measurement strategies
Figure 4. Texture identification of the tactile sensor assisted by spectral analysis and deep learning
Figure 5. Application in medicine of clinical feature identification based on the dual - mode tactile sensor and the interaction interface
Figure 6. Agricultural application of the dual - mode tactile sensor in intelligent white - strawberry picking
Original title: Quantitative softness and texture bimodal haptic sensors for robotic clinical feature identification and intelligent picking
First author of the paper: Ye QiuOriginal