Axle Counter Signal Normalization for Interference Rejection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Axle counting systems face interference from various disturbances, leading to incorrect detection of wheel passages, particularly in environments with tight curves or bogies, where noise and signal patterns can be misinterpreted.
Innovation Solution
The method employs amplitude and dynamic time normalization of measurement signals to standardize wheel passage patterns, allowing for reliable differentiation between genuine wheel signals and interference, using computer-aided evaluation to compare normalized signals with predefined patterns, including those for bogies and errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional axle counting methods are used, then the system is simple to operate, but interference signals cause incorrect detection of wheel passages
Solution Approach 1:
The patent applies preliminary action by normalizing the measurement signal before pattern comparison. The signal is pre-processed through amplitude normalization and dynamic time normalization, which standardizes the signal characteristics before the actual wheel passage detection occurs. This pre-processing step ensures that subsequent pattern matching operates on consistent, standardized data, improving detection reliability while managing complexity through systematic pre-computation.
Solution Approach 2:
The patent employs parameter changes by transforming the measurement signal through normalization processes. The amplitude is scaled to a standard range, and the time axis is dynamically adjusted to account for varying train speeds. These parameter transformations convert variable-speed, variable-amplitude signals into standardized forms that can be reliably compared against reference patterns, directly addressing the interference problem while maintaining manageable system complexity.
2Measurement precision
If amplitude normalization is applied, then the maximum signal amplitude is standardized, but the temporal profile varies causing detection errors
Solution Approach 1:
The patent applies segmentation by dividing the signal processing into distinct stages: first amplitude normalization is applied to standardize the maximum amplitude, then dynamic time normalization is applied to the temporal profile. This segmented approach allows each normalization type to address specific aspects of signal variability independently, with amplitude normalization handling magnitude consistency and time normalization handling duration variations due to speed changes.
Solution Approach 2:
The patent employs dynamics through dynamic time normalization, which adaptively adjusts the temporal profile of the measurement signal based on actual train speed variations. Unlike static time-windowing, this dynamic approach scales the time axis of the signal to match reference patterns, accounting for the fact that faster trains produce shorter-duration signals while slower trains produce longer-duration signals. This dynamic adjustment ensures consistent pattern matching across varying operating conditions.
3Productivity
If pattern comparison is performed without normalization, then the evaluation is computationally efficient, but interference signals are misinterpreted as wheel passages
Solution Approach 1:
The patent applies preliminary action by performing normalization computations before the pattern comparison stage. The amplitude normalization and dynamic time normalization are computed in advance, transforming the raw measurement signal into a standardized form that matches the reference patterns. This pre-processing ensures that when the actual pattern comparison occurs, it can be performed efficiently on already-normalized data, maintaining computational speed while improving detection accuracy.
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a method for counting axles, in which an axle counter sensor mounted on a track is passed by a wheel, the axle counter sensor generates a measurement signal (U1 ... U2), and the waveform (VL1 ... VL2) of the measurement signal (U1 ... U2) is evaluated by a computer to identify the wheel. During the evaluation of the measurement signal (U1 ... U2) within its waveform (VL1 ... VL2), at least one maximum (M1 ... M4) of the signal amplitude is sought. The amplitude of the measurement signal (U1 ... U2) is normalized by amplitude normalization such that the maximum (M1 ... M4) is identical to a predetermined target value (ZW). Dynamic time normalization is performed on the waveform (VL1 ... VL2) of the measurement signal (U1 ... U2) before and after the maximum (M1 ... M4). The normalized waveform (NV1 ... NV2) of the measurement signal (U1 ... U2), determined by amplitude normalization and time normalization, is matched with patterns (M1 ...M2) compares both at least one waveform (VL1 ... VL2) for the measurement signal (U1 ... U2) when passing a wheel (RD) and at least one waveform (VL1 ... VL2) for the measurement signal (U1 ... U2) when an error occurs. Furthermore, the invention comprises a computer program product and a provisioning device for the computer program product.