Anomaly detection method based on cloud edge fusion environment

An anomaly detection and integrated environment technology, applied in the field of the Internet of Things, can solve the problems of invalid data transmission, large energy consumption, and inaccurate detection results, and achieve the effects of expanding the scope, improving accuracy, and reducing resource and energy consumption

CN113840255AActive Publication Date: 2021-12-24HEBEI UNIV OF TECH
4 Cites 1 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Publication Date
2021-12-24

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention relates to an anomaly detection method based on a cloud edge fusion environment. The anomaly detection method comprises the following steps: step 1, screening m main attributes which have a decisive effect on an event; step 2, dividing a detection area into a plurality of sub-areas, and combining neighbor sub-areas to construct a spatial index tree; and querying the main attribute information of the event by using the spatial index tree, sending the main attribute information to the edge node, and performing abnormality judgment; step 3, performing anomaly detection of a single edge node; step 4, performing anomaly detection on the neighbor edge nodes of the abnormal edge nodes by using a social consciousness relationship; and 5, after the edge layer uploads the main attribute information of the abnormal sub-areas to a cloud layer, enabling the cloud to make a decision. According to the method, whether the neighbor edge nodes of the abnormal edge nodes are abnormal or not is judged through interaction existing in the social consciousness relation, the abnormal nodes can be guided to find the neighbor edge nodes which are possibly abnormal, the anomaly detection range is expanded, and the anomaly detection accuracy is further improved.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the technical field of the Internet of Things, and in particular relates to an abnormal detection method based on a cloud-edge fusion environment. Background technique

[0002] In order to ensure the safety of people's livelihood and network security, anomaly detection is particularly important in the Internet of Things. Anomaly detection in the Internet of Things environment has become an indispensable task for the development of social security. With the rapid development of Internet of Things technology, smart homes, wearable devices, and external devices for environmental monitoring all provide reliable technical support for anomaly detection. The development of edge computing has brought more convenience to time-critical anomaly detection tasks in the Internet of Things environment. Edge computing mainly processes data through the edge layer close to the data source to reduce the number of devices directly transmitting dat...

Examples

Embodiment

[0074] In this embodiment, the detection of indoor fire is taken as an example. The multi-attribute abnormal event detection method based on the cloud-edge fusion environment includes the following steps:

[0075] Step 1. Select a set of historical data for detecting indoor fires, and select data points from the historical data of detection equipment in a space area according to the time series at intervals of 1 min to obtain the original attribute information sequence; the attributes include temperature, smoke, CO, CO 2 , O 2 7, heat and watering amount, these 7 attributes are substituted into the multivariate logistic regression model of formula (1), and the regression coefficients of the 7 attributes are solved by maximum likelihood estimation, and temperature, smoke and CO are selected according to the regression coefficients from large to small The three attributes of concentration are used as the main attribute, that is, m=3;

[0076] Step 2. In order to make the edge n...