Application of Machine Learning to Include Honking Effect in Vehicular Traffic Noise Prediction

Random effect generalized linear model-based predictive modelling of traffic noise

Noise pollution is one of the negative consequences of growth and development in cities. Traffic noise pollution due to traffic growth is the main aspect that worsens city quality of life. Therefore, research around the world is being conducted to manage and reduce traffic noise. A number of traffic noise prediction models have been proposed employing fixed effect modelling approach considering each observation as independent; however, observations may have spatial and temporal correlations and unobserved heterogeneity. Random effect models overcome these problems. This study attempts to develop a random effect generalized linear model (REGLM) along with a machine learning random forest (RF) model to validate the results, concerning the parameters related to road, traffic and environmental conditions. Models were developed based on the experimental quantities in Delhi in year 2022-2023. Both the models performed comparably well in terms of coefficient of determination. Random forest models with R2= 0.75, whereas random effect generalized linear model had an R2= 0.70. REGLM model has the ability to quantify the effects of explanatory variables over traffic noise pollution and will be more helpful in prioritizing of resources and chalking out control strategies.

Heterogeneous road traffic noise modeling at mid-block sections of mid-sized city in India

This study attempted to develop a computer-based software for monitoring the traffic noise under heterogeneous traffic condition at the morning peak (MP), off peak (OP), and evening peak (EP) periods of mid-block sections of mid-sized city in India. Traffic noise dataset of 776 (LAeq, 1hr) were collected from 23 locations of Gorakhpur mid-sized city in the state of Uttar Pradesh in India. K-nearest neighbor (K-NN) algorithm was adopted for traffic noise prediction modeling. Moreover, principal component analysis (PCA) technique was used for the dimensionality reduction and to overcome the problem of multi-collinearity. The developed model exhibits R2 value of 0.81, 0.78, and 0.77 in the MP, OP, and EP, respectively, for Leq, and a value of 0.86, 0.80, and 0.84 for L10. The proposed model can predict more than 94% observations within an accuracy of ±3%. Ultimately, a user-friendly noise level calculator named "Traffic Noise Prediction Calculator for Heterogeneous Traffic (TNPC-H)" was developed for the benefit of field engineers and policy planners.

A hybrid deep leaning model for prediction and parametric sensitivity analysis of noise annoyance

Noise annoyance is recognized as an expression of physiological and psychological strain in acoustical environment. The studies on prediction of noise annoyance and parametric sensitivity analysis of factors affecting it have been rarely reported in India. A hybrid ConvLSTM technique was developed in the study to predict traffic-induced noise annoyance in 484 people based on ambient noise levels, as well as survey information. Ambient noise levels were obtained at different locations of Dhanbad city using sound level meter at varying intervals, viz. 09AM-12PM, 03PM-06PM, and 08PM-11PM. The proposed method was compared with some well-known neural network techniques such as K-nearest neighbors (KNN), artificial neural network (ANN), recurrent neural network (RNN), and long-short-term memory (LSTM). The experimental results indicate that the proposed method outperforms other techniques and can be a reliable approach for prediction of noise annoyance with an accuracy of 93.8%. It can be concluded from noise maps that the noise levels in all locations of the Dhanbad city were higher than 70 dB(A) and noise sensitivity is the most important input variable of traffic-induced noise annoyance, followed by honking noise, education, exposure hours, L_Aeq, sleeping disorder, and chronic disease. The study shall facilitate in developing a decision support tool for prediction of noise annoyance and promoting implementation of suitable public policy in urban cities.

Estimating traffic noise over a large urban area: An evaluation of methods

Unlike air pollution, traffic-related noise remains unregulated and has been under-studied despite evidence of its deleterious health impacts. To characterize population exposure to traffic noise, both acoustic-based numerical models and data-driven statistical approaches can generate estimates over large urban areas. The aim of this work is to formally compare the performances of the most common traffic noise models by evaluating their estimates for different categories of roads and validating them against a unique dataset of measured noise in Long Beach, California. Specifically, a statistical land use regression model, an extreme gradient boosting machine learning model (XGB), and three numerical/acoustic traffic noise models: the US Noise Model (FHWA-TNM2.5), a commercial noise model (CadnaA), and an open-source European model (Harmonoise) were optimized and compared. The results demonstrate that XGB and CadnaA were the most effective models for estimating traffic noise, and they are particularly adept at differentiating noise levels on different categories of road.

A Comparative Study of Ensemble Models for Predicting Road Traffic Congestion

Increased road traffic congestion is due to different factors, such as population and economic growth, in different cities globally. On the other hand, many households afford personal vehicles, contributing to the high volume of cars. The primary purpose of this study is to perform a comparative analysis of ensemble methods using road traffic congestion data. Ensemble methods are capable of enhancing the performance of weak classifiers. The comparative analysis was conducted using a real-world dataset and bagging, boosting, stacking and random forest ensemble models to compare the predictive performance of the methods. The ensemble prediction models are developed to predict road traffic congestion. The models are evaluated using the following performance metrics: accuracy, precision, recall, f1-score, and the misclassification cost viewed as a penalty for errors incurred during the classification process. The combination of AdaBoost with decision trees exhibited the best performance in terms of all performance metrics. Additionally, the results showed that the variables that included travel time, traffic volume, and average speed helped predict vehicle traffic flow on the roads. Thus, the model was developed to benefit transport planners, researchers, and transport stakeholders to allocate resources accordingly. Furthermore, adopting this model would benefit commuters and businesses in tandem with other interventions proffered by the transport authorities.

Optimized Sensors Network and Dynamical Maps for Monitoring Traffic Noise in a Large Urban Zone

We review a Dynamap European Life project whose main scope was the design, commissioning, and actual implementation of “real-time” acoustic maps in a district of the city of Milan (District 9, or Z9, composed of about 2000 road stretches), by employing a small number of noise monitoring stations within the urban zone. Dynamap is based on the idea of finding suitable sets of roads displaying similar daily traffic noise behavior, so that one can group them together into single dynamical noise maps. The Dynamap sensor network has been built upon twenty-four monitoring stations, which have been permanently installed in appropriate locations within the pilot zone Z9, by associating four sensors to each one of the six group of roads considered. In order to decide which road stretches belong to a group, a non-acoustic parameter is used, which is obtained from a traffic flow model of the city, developed and tested over the years by the “Enviroment, Mobility and Territory Agency” of Milan (EMTA). The fundamental predictive equation of Dynamap, for the local equivalent noise level at a given site, can be built by using real-time data provided by the monitoring sensors. In addition, the corresponding contributions of six static traffic noise maps, associated with the six group of roads, are required. The static noise maps can be calculated from the Cadna noise model, based on EMTA road traffic data referred to the ‘rush-hour’ (8:00–9:00 a.m.), when the road traffic flow is maximum and the model most accurate. A further analysis of road traffic noise measurements, performed over the whole city of Milan, has provided a more accurate description of road traffic noise behavior by using a clustering approach. It is found that essentially just two mean cluster hourly noise profiles are sufficient to represent the noise profile at any site location within the zone. In order words, one can use the 24 monitoring stations data to estimate the local noise variations at a single site in real time. The different steps in the construction of the network are described in detail, and several validation tests are presented in support of the Dynamap performance, leading to an overall error of about 3 dB. The present work ends with a discussion of how to improve the design of the network further, based on the calculation of the cross-correlations between monitoring stations’ noise data.