About this Journal Submit a Manuscript Table of Contents
Advances in Mechanical Engineering

Volume 2014 (2014), Article ID 616093, 12 pages

Research Article

New Regression Models for Predicting Noise Exposure in the Driver’s Compartment of Malaysian Army Three-Tonne Trucks

1Department of Mechanical and Material Engineering, Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia (UKM), 43600 Bangi, Selangor, Malaysia

2Science and Technology Research Institute for Defence (STRIDE), Ministry of Defence, Taman Bukit Mewah Fasa 9, 43000 Kajang, Selangor, Malaysia

3National Defence University of Malaysia (UPNM), Sungai Besi Camp, 57000 Kuala Lumpur, Malaysia

Received 31 August 2013; Revised 13 January 2014; Accepted 20 January 2014; Published 16 April 2014

Academic Editor: Francisco D. Denia

Copyright © 2014 Shamsul Akmar Ab Aziz et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


The objective of this study is to present a new method for determination of noise exposure in the driver’s compartment of Malaysian Army (MA) three-tonne trucks based on changing vehicle speed using regression models and the statistical analysis method known as Integrated Kurtosis-based Algorithm for -notch filter (I-kaz). The test was conducted on two different road conditions: tarmac and dirt roads. Noise exposure was measured using a sound level meter which is capable of recording raw sound pressure in Pa, and comparisons were made between the two types of roads. The prediction of noise exposure was done using the developed regression models and 3D graphic representations of the I-kaz coefficient . The results of the regression models show that increases when vehicle speed and noise exposure increase. For model validation, predicted and measured noise exposures were compared, and a relatively good agreement has been obtained between them. It was found that the predictions had high accuracies and low average relative errors. By using the regression models, we can easily predict noise exposure inside the truck driver’s compartment. The proposed models are efficient and can be extended to the automotive industry for noise exposure monitoring.

1. Introduction

Of late, issues regarding vibration and noise control in the driver and passenger compartments of vehicles have received increasing attention, with increasing awareness on noise, vibration and harshness (NVH) issues [1]. Driver and passenger performances in military vehicles are seriously affected by high noise levels, which can lead to hearing loss, degraded communication, misunderstandings, errors, accidents, and even failure to accomplish missions. Although these adverse effects may be minimal in stationary or slow moving vehicles, modern army tactics require rapid movement with constantly changing tactical situations [2]. High levels of noise exposure in high-speed moving vehicles can affect hearing and communication during operations. Researchers have identified impaired cognitive abilities attributed to driving at various noise levels, which include annoyance, speech and sleep interference, decreased work performance, hearing loss, and physiological changes [3, 4]. Furthermore, noise produced by cars can cause hypertension and can sometimes decrease driving focus, which may cause accidents [5].

MIL-STD-1474 (Noise Limits for Military Materiel) was drafted as the standard to provide noise limits for preventing hearing loss and on communication requirements for military personnel [6, 7]. The basic steady state noise criterion for conservation of hearing is that unprotected personnel should not be exposed to noise levels that are higher than 85 dB(A). Direct person-to-person communication is difficult when noise levels exceed this level [6]. Noise-induced hearing loss is one of the most common occupational disabilities among soldiers. Over 50% of US combat arms careerists developed significant hearing loss after 15 years of service [2]. The Malaysian Armed Forces (MAF) Medical Board reported that an increasing number of Malaysian Armed Forces (MAF) personnel and veterans are suffering from hearing problems, with 22% of them found to be suffering from the problem between 2000 and 2008 [8].

Various studies on noise model development have been conducted by many researchers in order to estimate or predict noise exposure from vehicles or road-traffic noise [5, 917]. Nopiah et al. [5] developed a vehicle acoustical comfort index (VACI) to evaluate the noise annoyance level and sound quality of noise. The test was conducted in two categories, stationary and highway, as the interaction between tyres and road surfaces gives major effect to the generated noise in the passenger compartment. It was concluded that the increase of engine speed can increase the annoyance level by decreasing the value of VACI; in other words it will contribute to more noise. Nor et al. [9] and Daruis et al. [10] evaluated VACI according to the most frequently used sound quality metrics, namely, Zwicker loudness, sharpness, roughness, and fluctuation of strength. As a result, the researchers could use this index for several types of roads and different road roughness characteristics. Rahmani et al. [11] and Gündoğdu et al. [12] modelled road-traffic noise using genetic algorithm (GA), which is one of the methods presently finding wide application in optimisation problems. The noise pollution level in the city was determined, followed by the development of a traffic noise level prediction model that can be used for the purpose of traffic noise level reduction by redesigning its flow or other means. They developed two models, which were time-independent (TID) that does not consider the direct effect of time and time-dependent (TD), that is, the modified form of the first model with consideration of the direct time effect. Suksaard et al. [13] and Pamanikabud and Vivitjinda [14] developed noise prediction models for environmental impact assessment in Thailand. Suksaard et al. [13] studied the average power levels of running vehicles, which were subsequently described by a relationship between power level and the logarithm of vehicle speed. Li et al. [15] developed an integrated geographic information system (GIS) based traffic noise prediction model based on data obtained from Beijing highways and neighbourhoods. This model produced results comparable in accuracy with those of the US Federal Highway Administration (FHWA) model that is already in use in China. The integrated noise-GIS system provides general functions for noise modelling as well as tools for noise abatement design. In addition, Pamanikabud et al. [16] developed a highway noise prediction model based on the equivalent sound level of over 20 s ( (20 s)). The use of (20 s), which is the real time measurement of average energy mean emission level of the entire noise path of 20 s of individual vehicles, can provide a more accurate measurement of individual vehicle noise. Meanwhile, Tansatcha et al. [17] used perpendicular propagation analysis to develop a motorway traffic noise model. The noise level that was generated by each type of vehicle was analysed according to the propagation in the direction perpendicular to the centre line of motorways carriageway.

Three-tonne trucks are commonly used vehicles for carrying personal and logistics in Malaysian Army (MA) services. The drivers of these trucks may be subjected to stresses, such as heat stress, vibration, air contaminants, and noise [18, 19]. Interior noise in the driver compartment is a combination of engine, road, intake and exhaust, aerodynamics, and noise from squeaks and rattles. Noise and vibration also originate from outside the vehicle, interacting with the vehicle structure and then producing radiated sound inside the driver compartment [18]. In previous studies [1, 20], it was demonstrated that as the speed of a three-tonne truck is increased, the noise level in the driver compartment increases. In this paper, the study is extended to present a new method for determination of noise exposure based on changing vehicle speed by using regression models and the statistical analysis method known as Integrated Kurtosis-based Algorithm for -notch filter (I-kaz).

2. Statistical Analysis Using Integrated Kurtosis-Based Algorithm for Z-Notch Filter (I-Kaz)

Statistical parameters, such as root mean square (RMS), kurtosis, crest factor, skewness, peak value, and signal-to-noise ratio (SNR), are widely used to detect differences in noise levels. While these indicators are easy to implement, the complexity of the mechanisms involved may give rise to serious errors in interpretation [21]. For this study, the investigation of statistical parameters derived from noise exposure time domain signals is performed using I-kaz. The statistical methods that summarise a collection of numerical or graphical data are called descriptive statistics, while the inferential statistical model is the pattern of data that permits randomness and uncertainty in the observations. I-kaz was developed for both descriptive and inferential statistics, whereby the numerical descriptor of I-kaz is the I-kaz coefficient , which is supported by three-dimensional graphical summarisations of the frequency distribution [22].

The sampling theorem states that perfect signal analysis is possible when the sampling frequency is greater than twice the maximum frequency of the sampled signal. Figliola and Beasley [23] suggested that the Nyquist number must be 2 or greater in order to avoid the content of the sampling signal from being misinterpreted. The maximum frequency span is described as follows:

The time domain signal is decomposed into three frequency ranges, which are -axis for low frequency (LF) range of 0–0.252 , -axis for high frequency (HF) range of 0.25–0.5 , and -axis for very high frequency (VF) range of higher than 0.5 . The selection of 0.25 and 0.5 as the low and high frequency limits, respectively, implies the concept of a 2nd order Daubechies in the signal decomposition process [22, 24]. In order to measure the scattering of the data distribution, the variance for each frequency band, which are , , and , is calculated, as shown in (2). The variance determines the average magnitude of deviation of instantaneous points with respect to the mean value:

The I-kaz method was developed based on the concept of data scattering about its centroid. The coefficient can be written in terms of the variance as follows: where , , , and , , and are the data for the th-sample of time and means of each axis, respectively, whereas is the numbers of samples in the data.

Equation (4) can be simplified in terms of kurtosis and standard deviation . Kurtosis, which is the signal of the 4th statistical moment, is a global signal statistic which is highly sensitive to the spikiness of the data. The Gaussian distribution of the kurtosis value is approximately 3.0. A higher kurtosis value indicates the presence of more extreme values than what should be normally found in a Gaussian distribution [25]. For discrete data sets, the kurtosis value is defined as Hence, (4) can be written in terms of and as follows where , , and and , , and are the and values of each axis, respectively.

The I-kaz method is capable of detecting small changes in the measured noise exposure. This method is used for extracting the raw data features of the sound pressure that was measured from the sound level meter during noise monitoring based on kurtosis, variances, and standard deviation. The recorded data is transferred to a computer and analysed by using MATLAB to calculate and produce the I-kaz display in real time. The I-kaz method is used to model the data patterns from the I-kaz display, which accounts for the randomness and draws inferences from a larger population. Therefore, these inferences are very useful for estimating and forecasting future observations [26].

3. Experimental Procedure

Noise in a MA three-tonne truck’s driving compartment was measured using a DuO smart noise monitor, with its calibration performed using a Brüel & Kjær 4231 calibrator before and after the measurements. This sound level meter is capable of recording raw instantaneous sound pressure in Pa for . The recording for duration of 180 s intervals contained 9,000 samples of instantaneous sound pressure raw data. The sampling period, which is the time difference between two consecutive samples, is 1/50 Hz = 0.02 s.

The measurement for each truck speed was repeated at least three times to ensure that the data is accurate and reliable for further analysis. All measured raw data was converted to sound pressure level (SPL) in dB and the arithmetic mean value for each trial was determined. The measurements were considered as valid if the range of the measurements for each road type and speed made immediately one after the other was not greater than 2 dB. The highest SPL value given by these measurements constituted the result [27] and was used as raw data for the I-kaz statistical analysis.

In addition to the evaluation of sound pressure data, the -weighted equivalent SPL was measured to determine the permissible exposure time for each truck speed. and sound pressure are related by the following equation: where is the reference sound pressure value and equals  Pa.

The microphone for this equipment was placed in the forward direction of the moving vehicle. The measurements were made at the same level as the ear position of the drivers as specified in ISO [28]. Figure 1 shows the dimensions used to measure the noise level at the driver’s ear position, with each microphone mounted at the centre line of a seat with height of 0.7 m.

Figure 1: Position of the microphone at the driver’s seat: (a) side and (b) plane views. Adapted from ISO [28].

Traffic and road situations and vehicle and environmental conditions offer unique challenges in utilising models for the purpose of noise exposure prediction [11]. Figure 2 shows the two kinds of road surfaces used for this study, which were tarmac and dirt roads (Table 1). The selection of different road surfaces was because interaction between truck tyres and road surface gives major effect to the generated noise in the truck driver’s compartment. The tarmac road has a flat, smooth surface and occasional unevenness, which resulted in minimum disturbances. The dirt road is an unpaved road made from subgrade materials and had frequent random irregularities which produced excessive casual vibrations. The tarmac road is the Kajang-Bangi highway with a two-lane highway along the side of Putrajaya, while the dirt road is a farmland in Port Dickson. Noise exposure was measured, while the vehicle was driven over the test roads.

Table 1: Characteristics of the road surfaces.
Figure 2: The two types of roads use for this study: (a) tarmac and (b) dirt roads.

In order to comply with ISO [27], the measurements had to be carried out under certain conditions. Both roads were chosen because the locations are quite far from main roads, to ensure that were no other noise sources during the measurements. Therefore, noise interference from others sources and vehicles could be reduced. The background noise (including wind noise) had to be at least 10 dB(A) less than the noise levels measured during the test. The background noise recorded using the sound level meter was in the range of 32 to 34 dB(A). The readings were repeated three times to get the average value for background noise. The low background noise indicated that it would not affect the noise measurement in the driver’s compartment. The truck’s windscreens were closed to prevent noise interference from outside. In addition, measurements were not made in adverse weather conditions [27]. On the day of the tests, the weather was sunny with temperatures ranging from 32 to 34°C and both roads surfaces were dry.

The tests were conducted for two conditions, which were stationary, and moving at 20, 40, 60, 80, and 100 km/h over the tarmac road and at 10, 20, and 30 km/h over the dirt road. Malaysia’s Federal Subsidiary Legislation on Environmental Quality Act 1974 [27] states that vehicle speed shall be measured with instruments with an accuracy of 3% or better. The Australian Motor Vehicle Standards Act [29] requires that speedometers installed in vehicles to be used on roads throughout Australia should have accuracy of ±10% for all speeds above 40 km/h. Furthermore, Witte and Wilson [30] reported that Global Positioning System (GPS) is 10 times more accurate than a vehicle speedometer for determination of speed over ground when moving at relatively constant speed in straight lines and is more accurate in determining speed on curved paths. Therefore, the truck’s speed was determined simultaneously using the speedometer fitted on the driver panel and a Garmin Nuvi 50 LM GPS receiver. The GPS receiver was mounted on the windscreen in front of the driver, so that the driver could maintain the truck speed measured on both devices during the measurement duration. The sound level meter reading was started when the speeds recorded on both devices were exactly the same.

4. Results and Discussion

4.1. I-Kaz 3D Graphical Representations of Time Domain Data

In order to quantify the overall impact of noise inside the driver’s compartment, the characteristics of the noise in the time and frequency domains are important. Standard diagnoses of the frequency domain are inadequate for evaluating the total effects of noise exposure in the driver’s compartment in the driving conditions. In order to evaluate the mutual correlation characteristics and detect any changes using a simple procedure in the actual complex working environment, it is necessary to introduce some signal processing methods, especially in the time domain [31].

Figure 3 shows a typical time history of the noise level inside the driver’s compartment for the two different types of roads. In this case, for every 0.02 s was recorded when the vehicle was travelling at 20 km/h over a distance of around 1 km. The highest value of (71.2 dB(A)) for the dirt road was higher than the tarmac road (70.9 dB(A)). The rough surface of the dirt road as compared to the smooth surface of the tarmac road caused the differences in noise levels inside the driver’s compartment. On the whole, the noise level inside the driver’s compartment running at relatively constant speed varied significantly with time.

Figure 3: -weighted SPL in the driver’s compartment when driving at 20 km/h: (a) tarmac and (b) dirt roads.

Using the I-kaz method, the signals obtained can be used to generate 3D graphical representations that show the values of . In Figures 4 and 5, the -, -, and -axes are the LF, HF, and VF ranges for tarmac and dirt roads, respectively. The figures show that for both types of roads, when vehicle speed is increased, the range values of become larger.

Figure 4: 3D graphical representations that show the values of for driving on the tarmac road at speeds of (a) 20 km/h, = 6.74 × 10−7, (b) 40 km/h, = 8.62 × 10−7 , (c) 60 km/h, = 1.15 × 10−6, and (d) 80 km/h, = 1.17 × 10−6.
Figure 5: 3D graphical representations that show the values of while driving on the dirt road at speeds of (a) 0 km/h, = 3.39 × 10−8, (b) 10 km/h, = 4.62 × 10−8, (c) 20 km/h, = 1.92 × 10−7, and (d) 30 km/h, = 4.10 × 10−7.

In Figure 4, for driving on the tarmac road, vehicle speed of 80 km/h produced the highest value of at 1.17 × 10−6. This is followed by 60, 40, and 20 km/h at 1.15 × 10−6, 8.62 × 10−7, and , respectively. For driving on the dirt road, Figure 5 shows when the truck’s speed is increased from 0 km/h to 30 km/h, also increases proportionately. The truck at idle position shows the lowest value of at 3.39 × 10−8. When the truck’s speed is increased from idle to 30 km/h, also increases proportionately with the highest value of at 4.10 × 10−7. This finding is strongly supported by the observation of the spaces of scatterings of the I-kaz displays in Figures 4 and 5. These graphical representations show bigger spaces of scatterings of values due to higher vehicle speed and noise exposure.

4.2. Mathematical Model for Predicting Noise Exposure

The mean values of noise exposure levels with 95% confidence limits were selected for further analysis. The results obtained for the tarmac and dirt roads are shown in Tables 2 and 3, respectively. The results for sound pressure and values are in agreement with Aziz et al. [1, 20], whereby higher vehicle speeds result in higher noise exposure. The curve fit method with linear polynomial regression was then used to plot the graphs for the relationships between and for tarmac and dirt roads.

Table 2: Values of sound pressure, , and for the various speeds of the truck while driving on the tarmac road.
Table 3: Values of sound pressure, , and for the various speeds of the truck while driving on the dirt road.
4.2.1. Tarmac Road

For driving on the tarmac road, Figure 6 shows the linear regression graph between and , with of 83.6%. The form of fitting equation is a linear polynomial equation as follows: where is the value of , is the value of , and and are constant coefficients which depend on and . Therefore, the developed equation for SPL predictions for driving on the tarmac road is as follows:

Figure 6: Linear regression graph between and for driving on the tarmac road.
4.2.2. Dirt Road

For driving on the dirt road, Figure 7 shows the linear regression graph between and , with of 97.0%. Based on the graph, the fitting equation in (8) was used to develop the equation for SPL predictions while driving on the dirt road, which is as follows:

Figure 7: Linear regression graph between and for driving on the dirt road.
4.2.3. Comparison of the Mathematic Models for Tarmac and Dirt Roads

Comparison of the mathematic models for driving on tarmac (9) and dirt (10) roads shows that there are differences in values of for predicting noise exposure inside the driver’s compartment, which are in the range of 3.39 × 10−8 to 1.17 × 10−6. The value of depends on the test condition’s parameters, which in the case of this study are vehicle speed and noise exposure. By inserting the value into the equations, we can get the noise exposure inside the driver’s compartment. For example, at the idle condition (  km/h), when the value of is , the value for both types of roads is 67.7 dB(A). The value is same because as the truck is idle, no extra additional noise is generated, other than the original noise generated inside the driver’s compartment. When the truck starts to move, also changes, with the highest value of at 1.17 × 10−6 producing value of 81.9 dB(A) for the tarmac road and at 4.10 × 10−7 producing value of  74.5 dB(A) for the dirt road.

4.3. Model Validation

In order to test the accuracy of the noise exposure prediction models developed in this study, three experimental sets for 40 datasets of noise exposure from different three-tonne trucks for each road surfaces were tested. This was conducted to evaluate the capability of the developed regression models to predict noise exposure inside the driver’s compartment. The predicted noise exposure was obtained using the equations that were developed for tarmac (9) and dirt (10) roads. The relative errors between the predicted and measured noise exposures were calculated using (11). The measured and predicted values of noise exposures and relative errors for the 40 datasets for the tarmac and dirt roads are shown in Tables 4 and 5, respectively: It can be seen from Tables 4 and 5 that the predicted noise exposures were very close to the measured noise exposures. The average relative errors were 1.35 and 1.45% for the tarmac and dirt roads, respectively, which are considered within the practically acceptable relative error limits of 10% [32]. Figure 8 shows the scatter diagrams of the predicted versus measured noise exposures for the 40 datasets for both types of roads. It is observed that the graphs fit very well and the correlations between the predicted and measured noise exposures follow the 45° line very closely. The predicted noise exposures have accuracies of 92.5 and 90% for the tarmac and dirt roads, respectively, which are within ±3 dB(A). The predicted results show better accuracy for the tarmac road, with smooth surface, as compared to the dirt road, with rough surface.

Table 4: Comparison of predicted and measured noise exposures for the tarmac road.
Table 5: Comparison of predicted and measured noise exposures for the dirt road.
Figure 8: Comparison between predicted and measured noise exposures for the (a) tarmac and (b) dirt roads.

5. Conclusions

This paper discussed the prediction of noise exposure inside the driver’s compartment of a MA three-tonne truck using regression models based on the I-kaz method. The inputs of the developed models are -weighted SPL , truck speed, and I-kaz coefficient . The I-kaz method was developed based on decomposed frequency signals to extract the features of sound pressure. It was found that sound pressure increase gradually with increasing vehicle speed and values. In addition, these results were supported by the observation of the scatterings of values in the I-kaz displays, which illustrated larger spaces of scatterings for increasing vehicle speed. It was found that that noise exposure is proportionally related to , with the accuracies of the developed regression models for the tarmac and dirt roads being 92.5 and 90%, respectively, which are within ±3 dB(A). It was also found that the predicted noise exposures generated by the regression models were close to the measured noise exposures with the low average relative errors of 1.35 and 1.45% for the tarmac and dirt roads, respectively. This indicates that the developed models can be used to predict noise exposure for the driver’s compartment of MA three-tonne trucks within acceptable limits. The performance of the predictions shows that the estimated results are very accurate and encouraging to be applied for noise exposure monitoring inside driver or passenger compartments of other vehicles.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.


The authors would like to thank the Government of Malaysia and Universiti Kebangsaan Malaysia for their financial support under INDUSTRI-2013-053 Grant. The authors are grateful to two anonymous reviewers for their suggestions that have helped strengthen this paper.


  1. S. A. A. Aziz, R. M. Sohaimi, M. H. Pu'ad, and M. A. M. Yaman, “Noise, Vibration and Harshness (NVH) study on Malaysian Armed Forces (MAF) tactical vehicle,” Applied Mechanics and Materials, vol. 165, pp. 165–169, 2012. View at Publisher · View at Google Scholar
  2. US Army Research Laboratory, Military Handbook: Design Guidance for Interior Noise Reduction in Light-Armored Tracked Vehicles, US Army Research Laboratory, 1993.
  3. M. M. Fooladi, “Involuntary and persistent environmental noise influences health and hearing in Beirut, Lebanon,” Journal of Environmental and Public Health, vol. 2012, Article ID 235618, 7 pages, 2012. View at Publisher · View at Google Scholar · View at Scopus
  4. M. D. Seidman and R. T. Standring, “Noise and quality of life,” International Journal of Environmental Research and Public Health, vol. 7, no. 10, pp. 3730–3738, 2010. View at Publisher · View at Google Scholar · View at Scopus
  5. Z. M. Nopiah, A. K. Junoh, W. Z. A. W. Muhamad, M. J. M. Nor, A. K. A. M. Ihsan, and M. H. Fouladi, “Linear programming: optimization of noise and vibration model in passenger car cabin,” International Journal of Soft Computing and Software Engineering, vol. 2, no. 1, pp. 1–13, 2012.
  6. Department of Defence US, Design Guidance for Interior Noise Reduction in Light-Armored Tracked Vehicles, vol. 767, Department of Defence US, 1993.
  7. Noise Limits for Military Materiel, MIL-STD-1474C, 1991.
  8. Bernama, More Armed Forces Staff have Hearing Problems, New Straits Times, 2011.
  9. M. J. M. Nor, M. H. Fouladi, H. Nahvi, and A. K. Ariffin, “Index for vehicle acoustical comfort inside a passenger car,” Applied Acoustics, vol. 69, no. 4, pp. 343–353, 2008. View at Publisher · View at Google Scholar · View at Scopus
  10. D. D. I. Daruis, M. J. M. Nor, B. M. Deros, and M. H. Fouladi, “Whole-body vibration and sound quality of Malaysian cars,” in Proceedings of the 9th Asia Pasific Industrial Engineering & Management Systems Conference, pp. 110–117, Bali, Indonesia, 2008.
  11. S. Rahmani, S. M. Mousavi, and M. J. Kamali, “Modeling of road-traffic noise with the use of genetic algorithm,” Applied Soft Computing, vol. 11, no. 1, pp. 1008–1013, 2011. View at Publisher · View at Google Scholar · View at Scopus
  12. Ö. Gündoğdu, M. Gökdağ, and F. Yüksel, “A traffic noise prediction method based on vehicle composition using genetic algorithms,” Applied Acoustics, vol. 66, no. 7, pp. 799–809, 2005. View at Publisher · View at Google Scholar
  13. T. Suksaard, P. Sukasem, S. M. Tabucanon, I. Aoi, K. Shirai, and H. Tanaka, “Road traffic noise prediction model in Thailand,” Applied Acoustics, vol. 58, no. 2, pp. 123–130, 1999. View at Publisher · View at Google Scholar · View at Scopus
  14. P. Pamanikabud and P. Vivitjinda, “Noise prediction for highways in Thailand,” Transportation Research D: Transport and Environment, vol. 7, no. 6, pp. 441–449, 2002. View at Publisher · View at Google Scholar · View at Scopus
  15. B. Li, S. Tao, R. W. Dawson, J. Cao, and K. Lam, “A GIS based road traffic noise prediction model,” Applied Acoustics, vol. 63, no. 6, pp. 679–691, 2002. View at Publisher · View at Google Scholar · View at Scopus
  16. P. Pamanikabud, M. Tansatcha, and A. L. Brown, “Development of a highway noise prediction model using an Leq20 s measure of basic vehicular noise,” Journal of Sound and Vibration, vol. 316, no. 1–5, pp. 317–330, 2008. View at Publisher · View at Google Scholar · View at Scopus
  17. M. Tansatcha, P. Pamanikabud, A. L. Brown, and J. K. Affum, “Motorway noise modelling based on perpendicular propagation analysis of traffic noise,” Applied Acoustics, vol. 66, no. 10, pp. 1135–1150, 2005. View at Publisher · View at Google Scholar · View at Scopus
  18. M. Harrison, Vehicle Refinement: Controlling Noise and Vibration in Road Vehicles, Butterworth-Heinemann, Boston, Mass, USA, 2004.
  19. A. Aybek, H. A. Kamer, and S. Arslan, “Personal noise exposures of operators of agricultural tractors,” Applied Ergonomics, vol. 41, no. 2, pp. 274–281, 2010. View at Publisher · View at Google Scholar · View at Scopus
  20. S. A. A. Aziz, M. Z. Nuawi, M. J. M. Nor, and D. D. I. Daruis, “Study of Noise, Vibration and Harshness (NVH) for Malaysian Army (MA) 3-tonne trucks,” Applied Mechanics and Materials, vol. 471, pp. 74–80, 2013. View at Publisher · View at Google Scholar
  21. M. J. Crocker, J. Wiora, K. Sharp, C. H. Hansen, and F. Jacobsen, “Gear fault diagnosis using bispectrum analysis of active noise cancellation-based filtered sound and vibration signals,” International Journal of Acoustics and Vibration, vol. 18, no. 2, pp. 58–70, 2013.
  22. M. Z. Nuawi, M. J. M. Nor, N. Jamaludin, S. Abdullah, F. Lamin, and C. K. E. Nizwan, “Development of integrated kurtosis-based algorithm for Z-filter technique,” Journal of Applied Sciences, vol. 8, no. 8, pp. 1541–1547, 2008. View at Scopus
  23. R. S. Figliola and D. E. Beasley, Theory and Design for Mechanical Measurements, John Wiley & Sons, New York, NY, USA, 5th edition, 2010.
  24. I. Daubechies, Ten Lectures on Wavelets, SIAM, Philadelphia, Pa, USA, 1992.
  25. J. A. Ghani, M. Rizal, M. Z. Nuawi, C. H. Haron, and R. Ramli, “Statistical analysis for detection cutting tool wear based on regression model,” in Proceedings of the International MultiConference of Engineers and Computer Scientists (IMECS '10), vol. 3, pp. 1784–1788, Kowloon, Hong Kong, March 2010. View at Scopus
  26. M. Z. Nuawi, A. Shahrum, A. Shahrir, S. M. Haris, and A. Arifin, MATLAB: A Comprehensive Reference for Engineers, McGraw-Hill, Selangor, Malaysia, 2009.
  27. Federal Subsidiary Legistlation, “Environmental quality act 1974—environmental quality (Motor Vehicle Noise) regulations,” 1987.
  28. “Acoustics: measurement of inside motor vehicles,” ISO 5128, 1980.
  29. Vehicle Safety Standards, Vehicle Standard (Australian Design Rule 18/00—Instrumentation), vol. 1, 2006.
  30. T. H. Witte and A. M. Wilson, “Accuracy of non-differential GPS for the determination of speed over ground,” Journal of Biomechanics, vol. 37, no. 12, pp. 1891–1898, 2004. View at Publisher · View at Google Scholar · View at Scopus
  31. H. Masuike and A. Ikuta, “Statistical signal processing by using the higher-order correlation between sound and vibration and its application to fault detection of rotational machine,” Advances in Acoustics and Vibration, vol. 2008, Article ID 828562, 7 pages, 2008. View at Publisher · View at Google Scholar
  32. Z. Şen, A. Altunkaynak, and M. Özger, “Space-time interpolation by combining air pollution and meteorologic variables,” Pure and Applied Geophysics, vol. 163, no. 7, pp. 1435–1451, 2006. View at Publisher · View at Google Scholar · View at Scopus