Control methods in a supernatural flight simulator

This video presents the experimental setup of an immersive flight simulation system, which combines body tracking with a headmounted display. Users are hanging freely in a climbing harness in order to improve the impression to fly. The base system was created by a group of students during one semester as a bachelor project. A study revealed that a high degree of presence is achievable by the combination of body-posture based control and 3D visualization via a head-mounted display while hanging freely in a climbing gear. Furthermore the results of a study that compares two different steering methods are presented. Both control methods seem to be quite interesting and intuitively controllable but show individual preferences among the participants of the study. In our experiments very frequently strong vaction effects were reported when flying turns to one side. The setup very likely causes simulator sickness, especially if a participant was bad in controlling his flight. 1 SYSTEM DESCRIPTION Some researchers already tried to build virtual flight systems with head-mounted displays (HMD). We can find systems which combine a hang glider harness and a small crane to an Iron Man setup. Others tried to build a kind of chair, like ICAROS VR [6] or Birdly [5]. As a key feature, the designers integrated mechanical parts into the realization of the flight control. The supernatural flight simuFigure 1: Simulator in use. Two different control methods are implemented. lator, proposed in this paper, is a multi-modal system with a high degree of freedom to move the arms, legs and the head within the environment. For our experiments, we extended an existing set-up of hanging traverses in our lab with a motorized mounting system for humans, limited to payloads up to 100 kg. A TOF camera sensor [4] is utilized for body tracking of the participant. The visualization is realized with an HMD with integrated IMU [2]. A powerful ∗e-mail: [email protected] †e-mail: [email protected] ‡e-mail: [email protected] graphics workstation renders the 3D scene in real time depending on user’s input by body movements with only low latency. We decided to place the Kinect sensor on the ground pointing upwards (fig. 2) because the data from skeleton tracking were more reliable than from ceiling position. Figure 2: Hardware setup at the user side.