Multitouch Sensor Noise Separation Using GRU and CNN
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing touch screens face challenges in accurately detecting touch inputs due to noise interference from components like displays and power systems, especially in portable devices where adding shielding materials is undesirable.
Innovation Solution
The use of machine learning techniques, specifically a gated recurrent unit (GRU) stage and a convolutional neural network (CNN) stage, arranged in series to mitigate noise in touch data. The GRU removes noise caused by one component, while the CNN addresses noise from another component.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If shielding materials are added to reduce noise from display and power system components, then noise mitigation improves, but device complexity and weight increase
Solution Approach 1:
The patent replaces physical shielding materials with a computational approach using machine learning models (GRU and CNN neural networks) to filter noise from touch data. The system processes touch data through these neural network stages to separate noise components from actual touch signals, eliminating the need for additional physical shielding layers in the device structure.
2Object-affected harmful factors
If shielding materials are added to reduce noise from display and power system components, then noise mitigation improves, but device weight increases
Solution Approach 1:
The patent substitutes physical shielding materials with software-based noise filtering using gated recurrent units and convolutional neural networks. This computational method processes touch data to identify and remove noise components without requiring additional physical materials, thereby maintaining device weight at original levels while achieving effective noise mitigation.
3Object-affected harmful factors
If post-processing algorithms are used to remove noise from touch data, then noise mitigation improves, but measurement precision may be compromised
Solution Approach 1:
The patent segments the noise removal process into two distinct neural network stages: a GRU stage that processes temporal sequences of touch data to remove time-dependent noise, and a CNN stage that processes spatial patterns to remove frequency-dependent noise. This segmented approach allows each stage to specialize in specific noise types while preserving genuine touch signal characteristics, thereby maintaining measurement precision.
Solution Approach 2:
The patent implements feedback mechanisms where the output of the GRU stage serves as input to the CNN stage, creating a cascaded processing system. The neural networks are trained with feedback from labeled data to learn the distinction between noise and actual touch inputs, enabling the system to maintain high measurement precision while effectively removing noise through iterative refinement.
Data Source
AI summary
In some examples, touch data can include noise. Machine learning techniques, such as gated recurrent units and convolutional neural networks can be used to mitigate noise present in touch data. In some examples, a gated recurrent unit stage and a convolutional neural network stage can be arranged in series, such as by providing the output of the gated recurrent unit as input to the convolutional neural network. The gated recurrent unit can remove noise caused by a first component of the electronic device and the convolutional neural network can remove noise caused by a second component of the electronic device, for example. Thus, together, the gated recurrent unit and the convolutional neural network can remove or substantially reduce the noise in the touch data.