| In recent years,new retail formats have developed rapidly,the application of internet technology has made it easier for consumers to purchase goods,and "unmanned" services are also continuing to penetrate into our lives.This article researches on the theme of smart cargo box,and combines the embedded platform and edge computing host to develop the smart cargo box system.First,this article proposes an embedded edge computing host,which has a variety of peripheral access interfaces,can interact with cloud computing platforms for wireless data,and has the ability to perform local calculations and analysis.The host uses the dual Rockchip RK3399 chip as the core.The design implements rich peripheral expansion interfaces such as Gigabit Ethernet interface,HDMI interface,SMA antenna interface,USB interface,RS232 interface,and it has the interactive capability of singlemaster and multiple-slave architecture communication protocol.Secondly,this article proposes and implements a variety of sensor solutions suitable for new retail stores and unmanned shelves.The first is a pressure sensor shelf module used for commodity weight detection;the second is a capacitive sensor shelf module used to distinguish products of approximate weight;the third is a distance sensor module for passenger flow detection;the fourth is based on RFID product identification technology,an RFID tag shelf structure with electromagnetic shielding characteristics is proposed when the product is placed,and the user takes the product to release the shield to achieve the purpose of interactive identification.Finally,this paper uses the edge computing host to combine multiple sensors for comprehensive applications,and proposes smart boxes and retail smart identification systems in different application scenarios.The first is a box system based on an electronic scale,and different types of goods can be identified based on their weight.The second is to design a cargo box system based on the interaction of pressure and capacitance sensing for some products with similar quality but different material and volume in the application scene.A product detection algorithm based on twodimensional input feature vectors is proposed,combined with pressure and capacitance sensing data,which can distinguish products of different materials and sizes without using a camera.The third type is a product identification system based on reverse RFID.It has designed a shelf structure with electromagnetic shielding characteristics for the RFID tags attached to the goods,and the user can remove the shielding to achieve the purpose of identification by taking the goods.In addition,the distance sensor as an optional accessory is installed on any shelf to provide passenger flow detection services,which proves the feasibility of the system design in various commercial scenarios.The cargo box system and product identification system designed in this paper can be well applied in a variety of business scenarios to meet the needs of offline merchants to attract passenger flow and increase sales. |
