Use of back-propagation neural networks to predict both level and temporal-spectral composition of sound pressure in urban sound environments

Torija Martinez, AJ ORCID: https://orcid.org/0000-0002-5915-3736, Ruiz, DP and Ramos-Ridao, A 2012, 'Use of back-propagation neural networks to predict both level and temporal-spectral composition of sound pressure in urban sound environments' , Building and Environment, 52 , pp. 45-56.

[img] PDF - Published Version
Restricted to Repository staff only

Download (1MB) | Request a copy

Abstract

One of the main challenges of urban planning is to create soundscapes capable of providing inhabitants with a high quality of life. Urban planners need tools that enable them to approach the final goal of designing, planning, and assessing soundscapes in order to adapt them to the needs of the population. Nowadays, authorities have models for predicting the A-weighted equivalent sound-pressure level (LAeq). Nevertheless, it is necessary to analyze not only the (LAeq) parameter but also the temporal and spectral composition of the sound pressure in the soundscape considered. The problem of modelling and predicting environmental noise in urban settings is a complex and non-linear problem. Therefore, in the present study, a prediction model based on a back-propagation neural network to solve this problem is proposed and examined. This model (STACO model) is intended to predict the short-term (5-min integration period) level and temporal-spectral composition of the sound pressure of urban sonic environments. Here, it is shown that the proposed model yields a precise and accurate prediction. Moreover, the results in this work demonstrate the validity of generalization of the STACO model, being applicable not only for the situations/locations measured, but also for any situation/location of a medium-sized urban setting, with some prior adjustment. In summary, the prediction model proposed in this study may serve as a tool for the integration of acoustical variables in city planning.

Item Type: Article
Schools: Schools > School of Computing, Science and Engineering
Journal or Publication Title: Building and Environment
Publisher: Elsevier
ISSN: 0360-1323
Related URLs:
Funders: Consejería de Innovación, Ciencia y Empresa de la Junta de Andalucía
Depositing User: Dr Antonio J Torija Martinez
Date Deposited: 02 Dec 2019 14:06
Last Modified: 03 Jan 2020 16:01
URI: http://usir.salford.ac.uk/id/eprint/53228

Actions (login required)

Edit record (repository staff only) Edit record (repository staff only)

Downloads

Downloads per month over past year