Research On Internet Of Things Based Intelligent Dining Cabinet System And User Recommendation

With the rapid development and maturity of the fifth generation mobile communication technology,the perfect combination of the Internet of things and information technology,big data analysis,artificial intelligence algorithm will also make the Internet of things develop rapidly,and be applied to different fields such as industry,agriculture,public service,etc.,among which the scale and types of application based on the Internet of things catering service industry are also increasing.In recent years,the turnover of the Internet catering service platform in China has continuously achieved a double-digit high-speed growth.In this trend,many online catering service platforms are also rising rapidly,such as “meituan”,“koubei”,“elme”,etc.,and one of the main ways for Internet users to eat online is to place orders directly on the online catering service platform of the Internet,and the merchants receive orders to complete the menu After that,the food will be distributed to the designated courier,who will deliver the food to the designated users.In this process,there may be a variety of problems,such as the delay of the courier,the failure of the courier to meet the user on time,the failure to place the received food at the time and place specified by the courier.In order to solve the problem of "the last 100 meters" in the traditional distribution of goods,this paper proposes an intelligent dining cabinet system based on the Internet of things,which can realize the seller's storage of food and the user's self-service meal.The dining cabinet system is divided into storage system,mechanical system,main control circuit system and user terminal system.The storage system includes the rotating grid and cabinet door;the mechanical system realizes the rotation function of the storage grid through the closed-loop servo motor and reducer.The main control circuit system uses raspberry pie as the main control board to develop its peripheral hardware control circuit;the terminal system realizes the control function by designing the serial display screen interface and assigning different storage addresses to the touch keys and input boxes of the interface,and raspberry pie UART serial port reads the feedback information of the display screen.The system is safe,convenient,responsive,professional and efficient,which improves the efficiency of intelligent sales service.While the development of Internet of things brings convenience,the problems of "information fragmentation" and "information overload" brought by user recommendation platform and big data application are increasingly prominent.Therefore,it is difficult for people to accurately find the recommended food that can meet their real needs and aspirations in this massive information.Although people can search on the platform recommendation website through a large number of keywords,they can not meet the user recommendation needs of big data personalization and user customization fragmentation.Therefore,personalized user recommendation algorithm emerges as the times require,and becomes a mainstream and solution to the problem of "information overload" user recommendation.This paper studies and compares the traditional text collaborative user filtering recommendation algorithm and the fragmented neural network user recommendation algorithm,and finds that the traditional text convolution and neural network based user recommendation algorithm has high efficiency,less mean square error(MSE),and more accurate and humanized recommendation results.