Multi-sensor anomaly detection and low power transmission method, system and storage medium

Abstract

The application provides a multi-sensor anomaly detection and low-power transmission method, system and storage medium. The transmission method comprises the following steps: maintaining a circular buffer on an edge device; performing statistical anomaly detection; performing calculation of multiple cross-sensor deviation degrees based on physical coupling relationship of sensors; counting the number of deviation degrees exceeding a deviation threshold, and processing in descending order of the number; generating an anomaly summary; forming a classification data stream and transmitting it to a server. The application improves anomaly distinguishing ability and traceability; reduces server-side analysis, storage and calculation load through edge-side filtering and summary reporting; raw sensor data is processed locally on the edge device, only statistical summaries and anomaly information are uploaded, which guarantees data security from the technical architecture level; unnecessary on-site maintenance is avoided by reducing false positives, battery replacement frequency is reduced by prolonging endurance, and the economy and reliability of the industrial monitoring system are improved.

Images