Peak Tracking Algorithm for Flat Object Boundary Layer Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for inspecting flat objects, such as wafers in the semiconductor industry, face challenges in accurately and quickly tracking and assigning peak signals in one-dimensional signal spaces, particularly due to interference and noise, which complicates the measurement of layer thicknesses and profiles.
Innovation Solution
An optical measuring system using spectral interference generates peak signals that are tracked and classified using a real-time capable peak tracking algorithm, employing a Kalman filter to estimate peak signal changes and assign them to specific types, enabling robust movement tracking even in complex situations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional object tracking methods using Kalman filters and multiple sensors are applied, then tracking capability is improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent extracts and utilizes only the essential motion information (peak position changes in one-dimensional signal space) needed for tracking, discarding unnecessary complex image data and multiple sensor requirements. This selective extraction maintains tracking reliability while significantly reducing system complexity.
Solution Approach 2:
The patent introduces a simplified peak tracking algorithm that acts as an intermediary between raw peak signals and classification results. This algorithm uses basic motion models to predict peak positions and assign peak types without requiring complex Kalman filters or multiple sensors, thus reducing system complexity while maintaining reliability.
2Measurement precision
If peak tracking is performed without motion prediction, then algorithm simplicity is maintained, but measurement precision and reliability deteriorate due to noise and interference
Solution Approach 1:
The patent applies preliminary motion prediction using simple motion models before peak classification. By predicting where peaks should appear based on their previous positions and estimated velocities, the system can accurately identify and classify peaks even in noisy conditions, improving measurement precision with minimal algorithmic complexity.
3Productivity
If rapid inspection of multiple locations is performed, then productivity increases, but measurement precision may deteriorate due to reduced measurement time per location
Solution Approach 1:
The patent implements continuous peak tracking across multiple measurement locations using motion prediction. As the measurement head moves continuously across the wafer, the algorithm continuously predicts peak positions and updates classifications without interruption. This continuous tracking maintains high productivity while preserving measurement precision through predictive correction of peak positions.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for precise and rapid peak tracking and classification, improving measurement accuracy and efficiency by reducing noise sensitivity and enabling continuous updates of peak positions, thus facilitating high-yield chip production and other industrial applications.
Implementation Method 1
The measuring system operates according to the method of spectral interference. It provides peak signals whose position in a one-dimensional signal space corresponds to the distance between the sensor and a boundary layer of the object or the distance between boundary layers of the object.
Data Source
Figure 1a~1b
Figure 2a
Figure 2b~2c
AI summary
A device for inspecting flat objects and detecting their boundary layers comprises an optical measuring system for measuring distances with a sensor. The measuring system provides peak signals whose position in a one-dimensional signal space corresponds to the distance between the sensor and a boundary layer or between boundary layers. The device also includes an evaluation unit configured to evaluate the peak signals in a tracking process and assign a predefined peak type to each peak signal. The evaluation unit assigns peak signals detected during a measurement to different tracks. Each track is updated with the assigned new peak signal. All peak signals within a track are assigned the same peak type. The assignment of a peak signal to a track is based on the change in the peak signal's position within the one-dimensional signal space.The change in a peak signal is estimated for the next measurement using an estimator.