Intelligent Emotion Detection Method in Mobile Edge Computing Networks

How to detect emotions of different people in time is significant in various areas, such as healthcare, VR/AR and etc. In this paper, an intelligent emotion detection method is proposed to transmit individual emotion data in real-time with as little energy as possible in a resource-limited mobile edge computing (MEC) networks. First, a data compression method is designed based on Convolutional Auto-Encoder (CAE) to compress emotion data on the user side. Then, to guarantee the requirements of data transmission delay and the data distortion rate simultaneously, an optimal channel allocation algorithm is proposed. Thereafter, the emotion data is recovered and analyzed by the edge personal emotion model EEGNET on the edge computing server where the real-time emotions of users can be monitored. The effectiveness of the proposed method is finally verified by extensive simulation experiments.