Fine-Grained Traffic Noise Prediction Using Segmented Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional traffic noise prediction models are computationally intensive and produce averaged noise levels that do not account for variations over time, making them inefficient and inaccurate for large areas.
Innovation Solution
Developed computational models that can be executed on consumer devices, predicting fine-grained traffic noise for specific times and locations, using vehicle count and class mix models to calculate noise levels, reducing resource usage and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional traffic noise prediction models are used, then comprehensive noise level assessment is achieved, but computational resources are excessively consumed and results are averaged over time losing temporal granularity
Solution Approach 1:
The patent segments the traffic noise prediction task into distinct components: vehicle count prediction, vehicle class mix prediction, and noise level calculation. Each component is handled by a separate model or calculation module, allowing for optimized resource usage at each stage while maintaining overall prediction accuracy.
Solution Approach 2:
The patent changes the temporal parameter from averaged noise levels to time-specific noise levels by incorporating temporal variables into the prediction models. This allows the system to provide fine-grained noise predictions for specific times while using efficient algorithms that reduce computational burden compared to conventional continuous simulation approaches.
2Loss of information
If conventional traffic noise prediction models are used, then noise level estimation is provided, but temporal variations in traffic noise are not captured
Solution Approach 1:
The patent introduces dynamic temporal parameters into the noise prediction model, allowing noise levels to vary over time based on traffic patterns. The system uses time-specific input data (such as hour of day, day of week) to dynamically adjust predictions, capturing temporal variations without requiring continuous simulation that would reduce productivity.
3Measurement precision
If detailed vehicle class mix prediction is implemented, then accurate noise level calculation is achieved, but model complexity increases
Solution Approach 1:
The patent segments the vehicle classification task into a dedicated vehicle class mix prediction model that outputs proportional distributions for different vehicle types. This separate module handles the complexity of class differentiation, while the main noise calculation module receives simplified input in the form of vehicle counts per class, reducing overall system complexity.
Solution Approach 2:
The patent applies different levels of detail to different vehicle classes based on their noise characteristics. The model focuses computational effort on distinguishing vehicle classes that have significantly different noise profiles, while using coarser classifications for vehicle types with similar acoustic signatures, thereby reducing model complexity while maintaining noise prediction accuracy.
Data Source
AI summary
A method for fine-grained traffic noise prediction includes obtaining location data indicating one or more locations and, for each respective location of the one or more locations, location-specific traffic data. The method includes, for each respective location of the one or more locations, using various computational models to predict a vehicle count and a vehicle class mix for the respective location; calculating, based on the predicted vehicle count and the predicted vehicle class mix, a number of vehicles per vehicle class for the respective location; and calculating, based on the number of vehicles per vehicle class, a predicted noise level for the respective location. The method includes causing a client device to display a map. The map may include, for each location of the one or more locations, a visual indication of the predicted noise level for the respective location.