Research And Realization Of Evaluation Service Of Remote FPGA Virtual Experiment Platform

For engineering courses,experimental courses occupies an important position.Traditional offline experimental teaching methods have problems such as unbalanced software and hardware resources,time and space constraints,so domestic and foreign began to explore online experimental modes.Our school's remote FPGA virtual experimental platform is also the result of online experimental exploration.By consulting related literature,it is found that the experimental data of many hardware remote experimental platforms at home and abroad are obtained by software simulation,which cannot reflect the real experimental situation,cannot combined with real experimental data for assessment.Based on the analysis and research of the remote FPGA virtual experiment platform,this topic combines the current more advanced technology framework to develop the evaluation function for the remote virtual experiment platform,and modularize the evaluation function.The assessment function,as an assessment auxiliary tool of the remote FPGA virtual experiment platform,makes up for the shortcomings of traditional assessment methods,improves the assessment efficiency of teachers,and is of great significance to the use of the platform and the learning effect of experimental courses.The main work of this paper includes the following aspects:First,research and design remote virtual experiment platform assessment micro services.According to the actual evaluation requirements of the platform,combined with the real data interacted with the remote experiment board,multiple evaluation methods are designed for result testing,auxiliary manual inspection,and record playback.For different types of experiments,the evaluation methods are dynamically selected to reflect the evaluation of hardware circuit experiments scalability and diversity.Second,develop file standards for assessment data based on the result test assessment method,and use the "asynchronous processing + Active MQ" method to process concurrent assessment requests to reduce implementation complexity and ensure platform response speed.According to the different demand scenarios of the experiment board,a variety of experiment board allocation strategies are designed to improve the use efficiency of the remote hardware experiment board and reduce the waiting time of users.Third,using Spring Boot's AOP programming ideas,while keeping the existing business processes and logic unchanged,use the learning behavior data that face the user to use the platform to record,and the user's learning behavior data and assessment results Carry out analysis and modeling,establish learning behavior evaluation model and performance prediction multiple linear regression model to help teachers improve teaching effects.Finally,the assessment services and related functions of the entire platform are jointly tested and verified,and they are continuously optimized during actual use,so as to provide strong support and guarantee for improving the efficiency of teacher assessment.