Research on Cooperative Control of Human-Computer Interaction Tools with High Recognition Rate Based on Neural Network

摘要

Human-Computer Interaction (HCI) is the vital technology of Virtual Reality. The control using several different HCI tools together provides a natural and effective method of human-computer interaction. This paper provides a way of cooperative control with multi-recognition tools. The separate tools, including speech recognition tools, hand gesture recognition tools, posture recognition tools, are developed firstly. Then assign the cooperation program receives the recognition result sent from the three independent tool through UDP protocol. A neural network that fits the habits of the operator is trained. Process the three recognition results using the trained BP neural network. Higher recognition rate compared with a single HCI tool is accomplished. Operators can communicate with the machine much more naturally and effectively.