Research And Development Of Intelligent Evaluation System Based On Remote Teaching Experimental Platform

The flow and liquid level measurement and control experimental devices generally occupy large areas and are expensive,which is difficult to meet the experimental teaching needs of the course Automatic Detection Technology in universities.Based on the flow and liquid level measurement and control experimental teaching platform of Southeast University,this paper develops an intelligent comprehensive experimental evaluation system.The system allows students to remote control devices and make experiments independently through the WEB webpage regardless of time and location.The platform intelligently schedules experimental devices and monitors the entire process of student experiments.It also can record and analyze the experimental process operations,experimental data and experimental report to conduct an intelligent comprehensive evaluation,which can improve the efficiency of equipment use and the scientificity of the evaluation in remote experimental teaching.This system has a highly practical value.First of all,this paper proposed the overall architecture of low delay and high expansibility system based on Ali Cloud.A remote experiment teaching platform based on the flow/liquid level integrated experiment device,remote online experiment integrated service system,and intelligent evaluation subsystem was established.Our web front-end realizes remote experiment remote control operation.And the back-end provided various services,like equipment scheduling,experiment process management and intelligent comprehensive evaluation.Secondly,in view of the characteristics of multiple experimental evaluation links and weak correlation,this paper proposes a set of intelligent multi-factor autonomous evaluation algorithms for remote measurement and control teaching experiments in order to enhance the scientificity of experimental evaluation.The autonomous evaluation system includes the experimental operation evaluation model based on the asynchronous cascade multi-task LSTM algorithm,experimental data evaluation method based on Ransac algorithm,experimental report evaluation algorithm based on text semantic analysis technology of BERT model and overall evaluation model based on multi-factor subjective weighting method.Finally,based on the B/S architecture and the RMVC design pattern,this paper builds a software platform for remote experiment intelligent evaluation.Include database,software and client system.The server design mainly includes the research and development of data realtime communication module based on socket.io,and the logical design of functional subsystems such as equipment monitoring and scheduling,online experiments,experiment evaluation and comprehensive services.A database system is built based on My SQL,and a reasonable and scalable data storage solution is designed.Based on the Express framework,module engine EJS technology,and third-party npm core modules,the corresponding client system has been developed to provide platform services for remote experiments and intelligent evaluation.The intelligent evaluation system of the remote flow and level measurement and control teaching experiment platform developed by this thesis has been officially launched in September 2019.After a full semester of operation testing,the system is stable and reliable,providing remote experiment service for nearly 150 students,about many design-type and comprehensive experiments were carried out on the platform.The novel experimental methods and evaluation methods effectively stimulated students' interest and achieved significant teaching effects.