Sensor Measurement Accuracy Estimation via Consensus Matrix
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for combining measurements from different sensors, such as those used in rail vehicles, face challenges in accurately determining speed due to measurement errors from slippage, tunnel interference, and satellite shadowing, leading to inaccurate speed readings and potential late arrivals at destinations.
Innovation Solution
A method that estimates the accuracy of different sensors by using a z-test to determine a consensus matrix, rescaling statistical normal distributions to increase consensus among sensor measurements, and weighting combined measurements based on their reliability, thereby filtering out outliers and ensuring reliable data fusion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If large error tolerances are assumed to accommodate slow-diverging measurements, then measurement reliability is improved, but measurement precision deteriorates
Solution Approach 1:
The patent applies dynamic adjustment of error tolerances based on the current measurement situation. Instead of using fixed large error tolerances, the system adapts the tolerance levels according to the consistency of measurements from different sensors, achieving both reliability and precision by making the tolerance criteria flexible rather than static
Solution Approach 2:
The patent implements a feedback mechanism where measurements from multiple sensors are continuously compared and evaluated for consistency. This feedback loop allows the system to identify when measurements are diverging slowly versus when they are outliers, and adjust the error tolerance assumptions accordingly, preventing both false rejection of valid measurements and acceptance of erroneous ones
2Measurement precision
If measured values from multiple sensors are combined with different weights, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent changes the parameter of measurement weighting from fixed predetermined values to dynamically determined values based on real-time consistency assessment. By using statistical methods to evaluate measurement consistency and automatically adjust weights, the system achieves higher precision without requiring complex manual configuration or adjustment mechanisms
Solution Approach 2:
The measurement system performs self-assessment of its own reliability by evaluating the consistency of measurements from different sensors. This self-service capability allows the system to automatically determine appropriate weighting factors without external intervention, reducing operational complexity while maintaining high measurement precision
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The invention relates to a method for estimating the measurement accuracy of different sensors (41, 42, 43) for the same measurand. In the method, measured values (m1, ..., mn) of the same measurand, which were captured by different sensors (41, 42, 43), are received. A consensus matrix (CM) is determined by applying a z-test having a predetermined significance value (p) on the basis of the captured measured values (m1, ..., mn) and an assumed statistical standard distribution (n1, ..., nn) of the measured values. The statistical standard distributions (n1, ..., nn) of the measured values (mi), which have the lowest consensus, are rescaled by the smallest scaling value (scmin), at which at least one of said measured values (mi) has an increased consensus. The last two steps are repeated until a desired consensus level is reached. The invention also relates to a measurement method for determining a measured value (m) of a measurand on the basis of measured values (m1, ..., mn) from a plurality of different sensors (41, 42, 43). In addition, the invention relates to an estimating device (30). Moreover, the invention relates to a measurement device (40). The invention also relates to a rail vehicle.