Supporting quality of service in real-time video streaming using the kinect

The field of networking is important as it allows us to communicate and exchange information amongst one another. However, in a real-time setting, it becomes important that such communication occurs with a high enough end-to-end quality to ensure a reasonable experience for all users involved. Specifically, when discussing how to stream 3D spatial data in real-time from a hardware sensor (such as the Kinect), there are numerous issues that arise. The hardware sensors used to collect the 3D spatial data generate them at a high rate and volume. The data rate (45 megabytes per second) is impractical to achieve, therefore we require additional techniques and solutions. In this thesis, we investigate the use of filtering, frame skipping, and compression within the context of the Point Cloud Library to alleviate the high data rate. The experiments conducted and the analyses generated suggest that, by utilizing these techniques, it is possible to support a quality of service in real-time video streaming using the Kinect.