Design And Implementation Of Embedded Real-time Database And Its Application In Anomaly Detection For Intelligent Agricultural Machinery

Intelligent agricultural machines have important economic sense to agricultural development.The real-time information system of intelligent agricultural machinery is used for the management of intelligent agricultural machinery equipment,with the real-time database as the data management platform,it stores and manages a large amount of real-time data such as component parameters,intermediate variables and control instructions during equipment operation.By connecting the configuration monitoring software it displays the system running state in real time,and by analyzing the historical data,the abnormal state detection of the system running can be realized.This paper does research into the key technologies of the real-time information system of intelligent agricultural machinery,and main work is as follows:(1)Design and implement the real-time database system.In the aspect of real-time database data management,the data structure of process data and measuring point data has been put forward,and design the data configuration and data preprocessing scheme.In term of data storage,index and cache mechanism are established for real-time data,and hierarchical storage design is used for historical data.In data compression,an improved swing door compression algorithm is proposed and its performance is verified.For data query,query schemes are designed for data stored in memory and disk respectively.(2)On the basis of the characteristics of anomaly data,an anomaly detection frame is proposed.Accessing historical data through the real-time database and based on time series analysis,through the training model in the normal sequence of samples,a multivariate gaussian distribution model is established for the errors of prediction model,anomaly detection is realized based on the prediction errors.A temporal convolution network is proposed for prediction model.The validity of the algorithm is verified by experimental evaluation on the data set of plant equipment running state.