Urban Noise Mapping via Context-Aware Matrix Factorization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise measurement systems in urban environments face challenges in accurately predicting noise levels and categories due to sparse data collection, which limits their ability to provide real-time and comprehensive noise maps.
Innovation Solution
A computer system that processes noise sample data using correlation information such as geographical and historical data to estimate noise levels in regions with sparse data, employing a context-aware matrix factorization approach to fill in missing data and provide real-time noise measurements across larger geographical areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If noise data is collected through crowdsourcing using mobile phones, then the coverage area and data volume increase, but the data sparsity problem worsens due to random and incomplete sampling
Solution Approach 1:
The patent introduces correlation information (geographic information, historical noise data, traffic data, weather data) as intermediary elements that mediate between sparse measurements and complete noise maps. These correlation data act as bridges to infer missing noise information in unsampled locations and time periods, resolving the contradiction between limited sampling and comprehensive coverage
Solution Approach 2:
The patent creates virtual copies of noise patterns by using correlation information to reconstruct and predict noise levels in locations and times where direct measurements are absent. Historical noise data serves as a template that can be copied and adapted to predict noise conditions under similar circumstances, filling gaps in the sparse measurement data
2Measurement precision
If traditional monitoring stations are used, then measurement accuracy is maintained, but the system complexity and deployment cost increase
Solution Approach 1:
The patent enables mobile phones to serve themselves as noise measurement devices by utilizing their built-in microphones and processors. The smartphones autonomously perform noise measurement, data tagging with GPS coordinates and timestamps, and wireless transmission to the central server, eliminating the need for dedicated monitoring station infrastructure and reducing deployment complexity
Solution Approach 2:
The patent transforms mobile phones from single-function devices into multi-functional tools that simultaneously perform noise measurement, location tracking, time stamping, and data communication. This universal utilization of existing mobile phone capabilities replaces specialized monitoring equipment, reducing system complexity while maintaining measurement functionality
3Area of stationary object
If more monitoring stations are deployed to improve coverage, then the noise map comprehensiveness increases, but the data processing and reconstruction complexity increases
Solution Approach 1:
The patent changes the fundamental parameters of data collection by shifting from fixed spatial monitoring stations to mobile, time-stamped measurements with GPS coordinates. This parameter transformation allows the system to handle sparse, random sampling patterns effectively using correlation-based reconstruction algorithms, reducing the complexity associated with managing dense networks of fixed monitoring stations
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer system measures noise in an urban environment using data records providing a location of a noise, a time stamp associated with the noise, and a noise category. Such data records are sparse with respect to both locations and time. The computer system also accesses other information that defines correlations of among different locations and among different time slots. Such correlation data can include geographic information and historical sample data. By applying this correlated information to the sparse data records about noise, the computer system can derive noise level and noise category information over a larger geographical area. Such information can be provided continually based on received data records about noise, typically in a matter of minutes after receiving the noise data for any given time slot.