基于多传感器的扶贫平台异常智能预警系统设计

Design of abnormal intelligent early warning system for poverty alleviation platform based on multi-sensor

扶贫平台中包含大量的数据信息,原有系统对异常数据界定不清晰,导致预警响应速度较慢,不能完全过滤掉异常数据,无法保证扶贫平台中工作项目的正常运行。为解决上述问题,设计基于多传感器的扶贫平台异常智能预警系统。在硬件设计中,构建扶贫平台异常分级架构,匹配多个智能预警传感设备,设计异常信号监测板电源电路。在软件设计上,基于多传感器构建异常数据融合模型,利用状态矩阵变换异常信号,均值化求解置信区域,发送异常数据的预警信号,完成智能预警系统设计。实验选择危房改造作为扶贫平台的测试项目,以错误数据和重复数据,作为扶贫平台的异常数据,测试该预警系统的响应速度和压力并发效果。在纯文字的错误信息预警过程中,该系统可保持在1.0 s以内,且能够过滤掉全部错误信息。在2 500组的新增用户中,设定100组重复数据,该系统可以做到全部重复信息的预警,具有实际应用价值。

The poverty alleviation platform contains a large amount of data information. The original system did not clearly define the abnormal data, resulting in a slow response to the early warning. The abnormal data cannot be completely filtered, and the normal operation of the work items in the poverty alleviation platform cannot be guaranteed. In order to solve the above problems, an abnormal intelligent early warning system based on multi-sensor poverty alleviation platform is designed. In the hardware design, build the anomaly classification architecture of the poverty alleviation platform, match multiple intelligent early warning sensing devices, and design the power circuit of the abnormal signal monitoring board. In software design, an abnormal data fusion model is built based on multi-sensor, abnormal signals are transformed by state matrix, confidence regions are solved by averaging, and warning signals of abnormal data are sent to complete the design of intelligent early warning system. The experiment selected dilapidated building reconstruction as the test item of the poverty alleviation platform, and used erroneous data and duplicate data as the abnormal data of the poverty alleviation platform to test the response speed and pressure concurrent effect of the early warning system. In the process of pure text error information warning, the system can keep within 1.0 s, and can filter out all error information. In 2500 groups of new users, set 100 groups of duplicate data. The system can alert all duplicate information, which has practical application value.