Design of Non-Intrusive Bus Data Monitoring and Intelligent Processing Unit

Huaizhi Li,Shengwei Meng,Shuying Dong,Peng Cheng,Yuchen Song
DOI: https://doi.org/10.1109/icemi52946.2021.9679563
2021-01-01
Abstract:State estimation and fault detection of complex equipment are of great importance to ensure operational reliability and safety. However, some key information is transferred between the sub-systems via different kinds of communication protocols. In other words, it is hard to measure or obtain the system state by additional sensors. On the other hand, limited by the bandwidth of the telemetry link, the real-time state estimation and monitor can only be implemented on-board. This brings more challenging issues for complex equipment state estimation, prognosis, and health management. Therefore, this work designs a non-intrusive bus data monitoring and intelligent processing unit (BMIP) for complex equipment online state estimation and anomaly detection. According to the communication protocol of the researched equipment, the design (BMIP) realizes the multi-channel 1553B and RS-485 communication. At the same time, the anomalies of the navigation system is online detected based on the least square support vector machine. The bus data is monitored and processed by an embedded system. Experiments have verified that the device has a good fault diagnosis result for constant discrepancy, drift and stuck. The use of this device can effectively reduce the cost of finding faults in industrial systems, and has a good application prospect.
What problem does this paper attempt to address?