Research on Big Data Consistency Algorithm of Multi-sensor Fusion

摘要

For the situation of data inconsistency and error irregular in multi-sensor detection, this paper based on the traditional measure of consistency, introduced Mahalanobis distance measurement to achieve approximate evaluation of data consistency between two sensors, then combined with the nature of membership function in the fuzzy theory to define support functions which measure the local measurement information and utilization efficiency, thereby obtaining assessments to sensor detection with better observations reliability, higher reliability. Simulation results show that, the weighted reliability to the measurement of multi-sensor observation data consistency can significantly improve the robustness of consistency measure, can eliminate the influence of sensor performance due to instability, effectively identify the sensor with unstable performance and remove it, providing data integration with sensors in consistency.