Road Traffic Noise Predictions by means of L10 Modelling with a Multilinear Regression Calibrated on Simulated Data

Rossi, Domenico; Mascolo, Aurora; Guarnaccia, Claudio

doi:10.46300/9104.2023.17.8

Cited by 1 publication

(1 citation statement)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For these reasons, the authors implemented a multiregressive technique for traffic noise assessment by calibrating it on computed data instead of real ones. As described in [36][37][38], such a regressive model has the advantage of not needing real data for its calibration. Moreover, the algorithms of generation of its calibration dataset make it potentially applicable to different traffic contexts.…”

Section: Introductionmentioning

confidence: 99%

Coupling Different Road Traffic Noise Models with a Multilinear Regressive Model: A Measurements-Independent Technique for Urban Road Traffic Noise Prediction

Rossi,

Pascale,

Mascolo

et al. 2024

Sensors

Self Cite

View full text Add to dashboard Cite

Road traffic noise is a severe environmental hazard, to which a growing number of dwellers are exposed in urban areas. The possibility to accurately assess traffic noise levels in a given area is thus, nowadays, quite important and, on many occasions, compelled by law. Such a procedure can be performed by measurements or by applying predictive Road Traffic Noise Models (RTNMs). Although the first approach is generally preferred, on-field measurement cannot always be easily conducted. RTNMs, on the contrary, use input information (amount of passing vehicles, category, speed, among others), usually collected by sensors, to provide an estimation of noise levels in a specific area. Several RTNMs have been implemented by different national institutions, adapting them to the local traffic conditions. However, the employment of RTNMs proves challenging due to both the lack of input data and the inherent complexity of the models (often composed of a Noise Emission Model–NEM and a sound propagation model). Therefore, this work aims to propose a methodology that allows an easy application of RTNMs, despite the availability of measured data for calibration. Four different NEMs were coupled with a sound propagation model, allowing the computation of equivalent continuous sound pressure levels on a dataset (composed of traffic flows, speeds, and source–receiver distance) randomly generated. Then, a Multilinear Regressive technique was applied to obtain manageable formulas for the models’ application. The goodness of the procedure was evaluated on a set of long-term traffic and noise data collected in a French site through several sensors, such as sound level meters, car counters, and speed detectors. Results show that the estimations provided by formulas coming from the Multilinear Regressions are quite close to field measurements (MAE between 1.60 and 2.64 dB(A)), confirming that the resulting models could be employed to forecast noise levels by integrating them into a network of traffic sensors.

show abstract

Section: Introductionmentioning

confidence: 99%