| Wearable devices are currently a popular direction,and their types and functions are also changing with each passing day.Smart bracelets and smart watches in daily life can be used for sign monitoring,smart helmets and smart glasses in the military can be used for target detection,virtual reality and other functions.These functions have high computational complexity and will give the endurance and user experience of the wearable device challenges.If other platforms can cooperate with the wearable device to perform calculations,reducing the amount of calculation of the wearable device will improve the performance of the wearable device.Because the cloud computing platform adopts a centralized processing model,the number of serviced devices is large and the distance to the user is far.The resulting network delay will make it difficult to meet the delay requirements of applications on wearable devices.With the rise of edge computing technology,the computing power of cloud computing can be transferred to the edge server close to the user at the edge of the network.The edge server can perform collaborative computing with the wearable device through the computing offloading technology to offload the application with high computing complexity on the wearable device to the edge server.This article first analyzes the application scenarios,functional requirements,and offloading requirements of the platform,and determines to use the application of the computing offload technology in edge computing to the wearing scenarios of team and team search.Based on the analysis of the existing computing offloading mechanism,a computing offloading mechanism based on Docker containers is established,and the compatibility of the computing offloading mechanism with the open source platform for edge computing is analyzed.The offload mechanism deploys an offload decision module on both the mobile terminal and the edge.The mobile offload decision module solves the problem of determining whether the application can be offloaded based on the energy consumption and time model of the wearable application,network status,and edge server load information.Based on the limited computing and communication resources of the edge server,the edge end offload decision module solves the application selection problem of different offloaded applications and the dynamic resource allocation problem of different container application services through greedy algorithms and mathematical optimization methods.Then based on the established computing offloading mechanism,a real edge computing offloading platform is built.Firstly,a combined communication method based on MQTT and HTTP is designed for the mobile terminal and the edge terminal.This method can make the request message transmit separately from the application execution data.The mobile terminal is mainly used to initiate an application offloading request,and an image management module,a container management module,a resource management module,and a data management module are designed on the edge.Finally,based on the motion recognition,face detection,target detection and data fusion of the four wearable applications in the scene,the platform's functions,offloading decisions and performance were tested.The experimental results show that the effectiveness of the offloading platform designed for team and team search scenarios. |
