Research And Implementation Of Cloud Robot System:CloudROS

The development of robots in the future is highly intelligent.The intelligent robots refer to the robots with highly developed "brains" of humanoids.The intelligent robots no longer perform repetitive operations in a step-by-step manner like traditional robots,but are able to constantly sense changes in the surrounding environment and react fastly.At the heart of intelligent robots is the ability of robots to process computationally intensive tasks quickly and efficiently.It requires the robot's onboard processor to equip more computing resources and storage resources.The cloud robot concept allows robots to acquire computing resources through the network,which offloads computationally intensive tasks of robots,and share information using the storage capability of the cloud infrastructure.Many cloud robot systems have been applied in robot tasks,but most cloud robot systems have a single adaptability,and there are problems such as difficulty in deployment,complicated data interfaces,and poor stability.ROS is suitable for almost all robot platforms,making it easy to design a unified data interface and implement distributed communication.In this pape,the cloud-based robot system based on ROS named CloudROS is presented.An independent ROS network runs on the robot,and the robot has the basic capabilty to complete tasks such as data collecting and motion control.In order to integrate cloud services into the ROS network,the cloud services are encapsulated in the form of ROS nodes,which make it possible to develop the robot applications and avoid data format conversion between the robot and the cloud server.The robots access the cloud services through the service interface provided by the cloud server,so that the cloud service nodes join to the ROS network running on the robots.The cloud server sets pre-processing stages including request authentication and service query to ensure smooth service scheduling,and introduces virtual isolation technology to ensure the security of service running.In order to ensure the stable running of services under dynamic network changes,this paper proposes a QoS regulation mechanism for cloud service quality.The mechanism integrates the impact of network environment and computational complexity to regularly monitor cloud service quality,and some reasonable dynamically adjusting measures that affect computational complexity are taken to ensure the stablity of cloud service under the dynamic network condition.In order to verify the effectiveness of the cloud robot system,the paper uses the Raspberry Pi,Arduino and other hardwares to design a robot mobile platform.The robot mobile platform is capable of performing basic motion controlling and environment sensing tasks.The experimental verification is divided into two parts:computing service verification and storage service verification.The robot transmits the image information collected by the binocular camera to the cloud processing to verify the effectiveness of the cloud robot system in unloading complex calculations.And the correctness of the QoS regulation mechanism is verified under the simulated dynamic network conditions.In addition,the multi-robot stores the collected robot motion information and environmental information in the cloud for centralized management,and the user monitors the state of the robot through the cloud,which verifies the storage sharing capability of the cloud robot system.