Design And Implement Of The Smart Lock Management System Based On Node.js

With the arrival of the era of 5G communications,the number of base stations of the Chinese three major telecom companys has added up to several million.Deploying smart locks for massive base stations and deploying corresponding management systems for station smart locks have become the tasks that the Chinese three major telecom companys have to complete in recent years.By analyzing the current research status of smart lock systems,we find that smart lock systems on the market are generally only suitable for homes and small businesses,and there is a shortage of large enterprise-class systems in the market,and many systems are not applicable to base station services.Many smart lock management systems use J2 EE or ASP.Net as the system architecture.The front and back end programming languages of these two architectures are inconsistent,which would increase the difficulty of system maintenance and the cost of human resources.Hence,it is extremely necessary to develop a smart lock management system based on suitable service architecture.Large enterprise-level system also has the high requirement for data encryption.The encryption algorithms of the large enterprise-level system should meet three requirements: Fast Processing Speed,High Security and encryption with Digital Signature Verification mechanism.And the large amount of business data of the large enterprise-level system has implicit business value,therefore,the lock smart system should use machine learning algorithm to do business information mining.Based on,a smart lock management system based on Node.js was designed and implemented.The main work of this paper is as follows:(1)The system designed by Node.js architecture and uses front and back end separation design,which conforms to the MVC separation principle and the Low coupling and High cohesion criteria,and the front-end and back-end system are developed with Javascript.Compared with the architecture of previous smart lock system,a unified development language makes the system development faster,easier to maintain and spend less money on maintenance.(2)The back-end server module is based on Node.js,Nginx and Mongdb technology,and the API interface of back-end server are developed based on POST protocol,the system has good platform scalability.(3)The front-end website module is designed based on the Html + javascript + css technology,and functions such as base station management,lock management,key management,organization management,work order management,staff management and data statistics are designed and implemented.Compared with the home-level or small company smart lock system,the system is more suitable for large enterprises with crossregional business.(4)A hybrid encryption algorithm based on Rijndael,RSA and SHA256 is designed.This algorithm has the advantages of fast symmetric encryption algorithm,asymmetric encryption security and has digital signature verification mechanism.Then use this algorithm for the communication layer.Compared with the encryption algorithm of previous smart lock system,the new encryption algorithm is superior in comprehensive performance.(5)A Fast Naive Bayesian Ensemble Learning Algorithm(Fast-NBE)is proposed.Compared with the existing Naive Bayesian Ensemble Learning Algorithm(NBE),the time complexity of training N samples and obtaining M basic classifiers is reduced from O(MN)to O(N + M).Finally,the API interface of business information mining module is designed based on Fast-NBE.Compared with the previous smart lock system,it is the first time to use machine learning algorithm for business information mining.In the past,many systems did not use machine learning algorithm or only uesd machine learning algorithm to do face recognition.In addition,Fast-NBE use the new ensemble output strategy based on logarithmic transformation and Softmax normalization(LC-SNO)to avoid the same number of votes in category output.