Machine learning based human interacted robotic intelligence to detect the different categories of speech

摘要

Realizing artificial intelligence depends heavily on human-robot interaction (HRI) technology. Here, we present dual-function sound actuators built on graphene for human-robot exchanges that use machine learning. (GHRI). The GHRI's triboelectric acoustic detection mechanism and thermoacoustic sound output mechanism allow it to serve as both a prosthetic ear and tongue. Triboelectric materials, electrodes, and thermoacoustic sources made possible by laser-induced graphene contributed to the overall success of the combined device. The GHRI can recognize speech identities, feelings, substance, and other information in the human speech by optimizing the structure parameters to produce high sensitivity and working longevity. With the help of machine learning, a convolutional neural network is used to teach 30 voice categories, with an accuracy of 99.72% and 97.83% in training and test datasets, respectively. In addition, AIs can use GHRI to communicate through recognized voice characteristics. The implications of our research for the future of artificial intelligence in robots are vast.