抄録
We propose a design method of virtual reality (VR) environment for maximizing VR presence and minimizing VR sickness by controlling angular velocity and visual angle and evaluate this method. The system that materializes the proposed design method has two neural networks. One neural network (NN) learns user's VR sickness characteristics. Another NN learns user's VR presence characteristics. The optimum condition for user is calculated by combining these two NNs. First, we analyze the overall tendency and the individual variation about the VR sickness and presence characteristic of subjects. As a result, it was confirmed that these two characteristics are in the trade-off relationship and that the individual variation is large. Next, we evaluate the effectiveness of our proposed design method. This method had a potential to design VR environment that realize low VR sickness and high VR presence taking account of the difference of user's characteristics. Especially, this method worked effectively for subjects who develop heavy symptom of VR sickness.
