Research on Abnormal Data Correction Methods for Remote Monitoring of Heavy-Duty Vehicles Oriented to Smart Supervision

Heavy-duty vehicles emissions are a serious problem, and remote monitoring platforms are a key means of emission control for heavy-duty vehicles. However, the frequent occurrence of anomalies in the remote monitoring data has seriously limited the monitoring efficiency of the remote monitoring platform. Therefore, this paper takes 500 National VI heavy-duty vehicles as the research object, and proposes a whole-process data quality control system of “anomaly identification-dynamic correction-accuracy verification”. First, four types of anomaly patterns, namely, lost, invalid, outlier and mutation, are defined, and polynomial fitting, median filtering and contextual interpolation are adopted to realize differentiated correction. Second, a data accuracy validation framework based on correlation analysis was constructed. The results show that the accuracy of key parameters is significantly improved after correction, and the data fitting degree R² is greater than 0.97. The research results ensure the accuracy of remote monitoring data and improve the regulatory efficiency of the platform, which is of great significance for the management and control of medium and heavy vehicles in intelligent transportation and green and low-carbon development.