Tactile-Olfactory Intelligent Sensing Arrays For Objects Recognition in Hazardous Environments

To date, visual-based detection is one of the most common methods used for object recognition. However, high recognition accuracy of visual-based detection mainly relies on high-quality cameras and substantial computing power. In comparison, other sensing strategies, such as tactile and olfactory, have relatively small input data size, smaller requirements of computing resource and faster identification speed, which are crucial in rescue mission. Here, we report a tactile-olfactory fusion system for object recognition by using tactile and olfactory based sensing arrays attached on a mechanical hand, which permits real-time acquisition of an object's local topography and odor. From the experimental results, the recognition rate after fusion is higher than sole perception; eleven different types of objects can be identified and classified with a recognition rate of 91.2%. Moreover, in burial scenarios, human and debris can be robustly identified with the recognition accuracy above 80%.