Time-Frequency Signal Processing for Multi-Sensor Noise Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods using multiple reference sensors for noise removal from desired signals often result in insufficient accuracy due to limitations in signal processing techniques.
Innovation Solution
An information processing device and system that acquires signal measurements from both signal and reference sensors, and performs signal processing by dividing the data into multiple frequency bands across time segments to effectively remove noise from the mixed signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional time-series filtering with multiple reference sensors is used, then the system can measure environmental noise, but the accuracy of noise removal from the desired signal is insufficient
Solution Approach 1:
The patent divides the time domain into multiple segments and further divides each segment into multiple frequency bands. This segmentation approach allows independent processing of different time-frequency regions, improving noise removal accuracy by adapting to local signal characteristics while maintaining manageable computational complexity through modular processing.
Solution Approach 2:
The patent applies different processing strategies to different segments and frequency bands based on local signal characteristics. By calculating correlation coefficients separately for each segment-frequency combination and applying adaptive filtering locally, the system achieves higher overall accuracy without uniformly increasing complexity across the entire signal.
2Adaptability or versatility
If the number of reference sensors is increased to handle multiple types of environmental noise, then the coverage of noise types improves, but the accuracy of noise removal becomes insufficient due to processing limitations
Solution Approach 1:
The patent segments the signal processing into multiple time segments and frequency bands, allowing each reference sensor's contribution to be optimized independently for different noise types in specific time-frequency regions. This enables effective utilization of multiple reference sensors for diverse noise types while achieving high removal accuracy through localized adaptive processing.
Solution Approach 2:
The patent dynamically adjusts processing parameters such as segment boundaries, frequency band divisions, and filtering coefficients based on the characteristics of each segment. This parameter adaptation allows the system to effectively handle multiple noise types with varying characteristics while maintaining high accuracy in each local region.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided is an information processing device capable of improving the accuracy of removing noise from a result of measuring a desired signal when a plurality of reference sensors are used. An information processing device includes an acquisition unit configured to acquire signal measurement results of one or more signal sensors for measuring a mixed signal in which an objective signal and noise are mixed and noise measurement results of a plurality of reference sensors for measuring the noise and a signal processing unit configured to divide the signal measurement results and the noise measurement results acquired by the acquisition unit into a plurality of frequency bands for each of a plurality of division segments in a time domain and perform signal processing for removing the noise included in the mixed signal.