Optimizing Camera Stream Transport in Cloud-Based Industrial Robotic Systems

摘要

Combining visual-guided robotics with cloud networking brought a new era into industrial robotic research and development. New challenges have to be tackled with a focus on providing proper communication and data processing setup: sensor data processing as well as the control software should be decoupled from the local robot hardware and should move into the cloud. In the emerging field of cloud robotics, there are trade-offs that have to be handled. More and more sensors such as cameras are being integrated but it comes with a cost. All sensory data have to be sent through often limited networking resources, while latency must be kept as low as possible. In this paper we propose a general solution for efficient camera stream transportation in cloud robotic systems. After introducing our test scenario with the used hardware and software elements, a detailed overview of the architecture is presented with describing each task of the components. The goal of this paper is to examine the current stream transportation implementations in ROS environment and implement a more efficient method. The performance of the proposed method is investigated and compared with other solutions evidenced by measurements.