Design And Implementation Of Virtual Robot System Based On Physics Engine And Augmented Reality

Virtual robot has a very wide range of applications in education and training,industrial production, simulation design and remote operation. It can reduce cost and provide users with a good display by modeling and simulating real devices. But it is rarely used for services and user interaction. In this dissertation we design and implement a virtual robot interaction system based on “SRU service robot” and augmented reality.It can demonstrate the SRU robot’s functions and movement. The system can save resources by demonstrating robot among various kinds of environments,and can reduce equipment wear, in addition it can replace part of the functions of the real robot(interactive function).Augmented reality is a kind of technology that to integrate 2D images or 3D virtual objects to real environment by calculating the camera’s position and angle real-timely by computer vision. It inherited the advantages of the virtual reality and made up for its shortcomings, and the display effect of AR is much more real than virtual reality.We will work in the following aspects:1. Calculate camera’s external parameters by camera calibration, and determine the position and angle of the camera. Analyze the storage structure of images,design and implement the transmission module.2. Design the layout of 3D scene, including camera, three-dimensional model, the light source, and model materials, and configure the properties of these components according to the real scene. Load physics engine in the virtual environment, configure the virtual robot’s physical properties according to SRU robot. Based on the above, we implement the virtual robot’s roaming function.In addition we design and implement data receiving module of the server,present a scene to users in which the virtual objects are integrated into real scene by updating the scene background image and camera’s position and angle real-timely.3. Generate high dynamic range environment mapping by collecting the light information, and convert high dynamic image to low dynamic image by using tone mapping method. Then convert spherical environment mapping to cube texture which is used to render the surface of virtual robot. In addition, we design the module that can adapt to ambient light dynamically.4. Design and implement the interactive module of system, it allows users to control virtual robot manually or specify the robot’s destination(by using ray positioning method).