System Design And Implementation Based On Decentralized Artificial Intelligence Internet Of Things

With the continuous innovation and improvement of the Internet of Things technology,the application of the Internet of Things has penetrated into various industries.Countries have included the development of the Internet of Things into their national-level strategies.According to IDC's data,in 2018,the global Io T spending reached 646 billion dollars.In the future,the global Io T market will continue to grow steadily.By 2022,the market size will reach 1 trillion dollars.The Internet of Things is an important part of the new generation of information technology and an important stage of development in the "informatization" era.It is known as the third wave of the world's information industry development after computers and the Internet.After Google launched Alpha Go in 2016,the development of artificial intelligence has grown exponentially.As technologies such as artificial intelligence and the Internet of Things continue to mature,artificial intelligence technology and Internet of Things technologies are combined to create a more intelligent and humanized Internet of Things to meet the increasing demands of people.AIo T is the integration of artificial intelligence technology and the Internet of Things in practical applications.At present,more and more people have combined AI and Io T.AIo T,as the best channel for intelligent upgrading of traditional industries,has become an inevitable trend in the development of the Internet of Things.AIo T is becoming the mainstream form of the future technology recognized in the industry,accelerating its application and landing worldwide.In the future,human production and life will be highly intelligent,entering an intelligent era.At present,the Io T system supporting artificial intelligence mainly sends all data collected by sensors to a central location,such as the cloud or a dedicated server with dedicated GPU hardware,where model building and inference tasks are performed.With the large increase of terminal equipment and the further improvement of people's requirements for service quality,the centralized processing mode centered on cloud computing has bandwidth,delay,and security problems.Traditional artificial intelligence Io T systems need to be always connected to the cloud,and must deal with bandwidth limitations and high latency variability.Depending on the application,the remotely processed and aggregated data is forwarded back to the actuator device that is usually located near the originating sensor node,and the cloud is often far from the underlying device.The Internet of Things system is connected with sensors,actuators and smart devices with huge data.Generally,actuators only need to respond to the data of all thedevices in the area,not all the devices.All the data is transmitted to the cloud,transmitting a large amount of inefficient data,Resulting in waste of network bandwidth.Sending sensor data to the cloud can introduce security holes and privacy issues.Because the communication path from the terminal to the cloud is long and there are many nodes,it is vulnerable to network attacks.In view of the above problems,this paper proposes a system based on the distributed artificial intelligence Internet of Things.This article introduces the idea of combining cloud and fog to provide computing and storage for the underlying terminal equipment at the edge of the network closer to the terminal equipment,and introduces Docker container virtualization technology in the system.The specific research contents and contributions of this article are as follows:(1)By analyzing some problems in the traditional artificial intelligence Io T system,a system based on decentralized artificial intelligence Io T is proposed.(2)The system based on decentralized artificial intelligence Io T is designed.According to the characteristics of BP neural network,a BP neural network with only one hidden layer is selected as the neural network model.The basic idea of the design is: one-to-one mapping of the Io T nodes and the neurons of the neural network,the sensing device in the Io T as the input neuron of the BP neural network,the fog device as the hidden neuron,and the execution device as the output neuron.The sensing device,the fog device and the execution device together form the BP neural network.The idea of combining cloud and fog is introduced in the architecture design.Set up decentralized and centralized working methods for the system.(3)The function of the system based on the distributed artificial intelligence Internet of Things is realized.Deploy Python and Docker on development devices,write Dockerfile files to build Docker private images,and build Docker containers based on Docker images.Deploy code in a Docker container to implement the functions of the fog device and the execution device,and use socket to implement communication between containers.(4)Aiming at the mapping problem between Io T nodes and neural network nodes,a system optimization mapping model was constructed.Specifically,the system optimization mapping model was constructed with the target functions of transmission time,transmission power,and the combination of transmission time and transmission power.