Design And Implementation Of Knowledge Centric Networking Prototype Architecture Based On Collaborative Computing

The rapid development of equipment miniaturization technology in these years has ushered in the application of distributed detection clusters in real-time target recognition and other scenarios.However,with the deepening of the application,the following problems arise.Restricted by the detection accuracy of a single node,a single node in a distributed detection cluster can only obtain low-feature dimension knowledge about the target when performing target recognition.Such knowledge often contains many interference targets.At the same time,due to the decentralized nature of the distributed detection cluster,the knowledge acquired by each node of the cluster will be directly uploaded to the upper-layer users.A large amount of knowledge containing interference targets is sent to upper-layer users at the same time,which will interfere with users'correct recognition and judgment of the targets.Aiming at these problems,some scholars have proposed a concept of a knowledge centric networking.The core design idea of the knowledge centric networking is to enhance the intelligence and interactivity of the system,improve the service quality,and communicate with better controllability and lower cost through the extraction and representation of knowledge,composition and evolution,effective upload and distribution.In order to avoid a large number of interference targets in the distributed detection cluster affecting the user's accurate identification of real targets,based on the design concept of the knowledge centric networking,this thesis studies and implements a knowledge centric networking based on collaborative computing with the goal of changing the amount of knowledge to the quality of knowledge.Firstly,this thesis studies a collaborative computing method for knowledge evolution,by which knowledge can be shared and evolved in distributed clusters.Then,based on collaborative computing method,a knowledge centric networking implementation scheme is studied and designed,and the scheme is tested and verified.The results show that the knowledge centric networking based on collaborative computing,which is studied and implemented in this thesis,can transform the knowledge with low feature dimension in distributed detection cluster into knowledge with high feature dimension and report the final high feature dimension knowledge to users.The design goal of changing the quantity of knowledge to knowledge quality in the distributed detection cluster is realized,and the influence of the interference target on the correct recognition of the user is avoided.