• cpx method
• profilometry
• road texture
• rolling noise
• tyre/road noise

Traffic noise is one of the main sources of acoustic pollution, with an ubiquitous presence in anthropised environments.Therefore, it is a constant component of residual noise during every noise measurement.
The generation and propagation of traffic noise is governed by three main different mechanisms: vehicles' power unit (i.e. the engine, gearbox and cooling system), vehicles' profile aerodynamics and tyre-road interaction.
While engine noise is the main source up to 35 km/h, for higher speeds propulsion noise is negligible over noise due to tyre/road interaction. For heavy vehicles, like lorries, the contribution of propulsion noise cannot be neglected even at higher speeds.
Tyre/road interaction depends greatly on the characteristics of road surfaces: a subtle difference in road surfaces may yield a large variation in noise levels. This idea has lead to a series of studies that focused on providing theoretical and phenomenological analysis of the influence of road profile on noise emission.
The characterisation of a road surface requires the measurement of its profile: the device used to carry out these measurements is known as profilometer; modern road profilometry is based on laser triangulation sensors, which allow a simple and direct measurement of road texture.
The purpose of this work, carried out at the CNR spin-off company iPool srl, was to develop a laser profilometer focused on measuring the macrotexture and megatexture of road pavements, which cover wavelengths of the same order of magnitude as the size of tyre tread elements and of the tyre/road interface and to verify some of the existing models to predict tyre rolling noise based on road texture knowledge. This work confirms the importance of macrotexture (MaTx) and part of megatexture (MeTx) in the generation of rolling noise. The profilometer, mounted on a vehicle, is composed of a laser displacement sensor, an accelerometer, known in jargon as inertial unit and a rotary encoder attached to a wheel, which provides the horizontal travelled distance. This profilometer presents many strong points, viz. a simple usage and relative low cost compared to other laser profilometers; moreover it presents a new procedure to obtain the road profile by combining the information provided by the laser output, a time-domain displacement signal and a record of the distance travelled by the vehicle, measured by the angular encoder.
Chapter 1 provides a description of the main terminology and devices used in profiling, while chapter 2 deals the essential theoretical background of the dominant phenomena involved in the generation of rolling noise and describes the main models that correlate rolling noise and road texture.
Chapter 3 accurately explains every component the profilometer; it is roughly divided in two main parts, the first one outlines the hardware components specifications and limits, while the second part overviews the signal processing. An algorithm in Matlab environment for data processing was also developed, dealing with the re-mapping of the time-domain vertical displacement signal with a constant sampling frequency to a space-domain profile, sampled every 0.5 mm, and subsequent evaluation of road texture parameters has been conducted on different sites. A simple mathematical model of road texture enveloped by the tyre has also been used during data analysis, simulating the actual road surface that is in contact with the tyre.
The following chapter addresses the possible sources of uncertainty of profiling operations. A theoretical model of the response of the laser sensor at different wavelengths has also been developed, taking into account the finite spot size of the laser and the finite exposure time of the pixel array. These two factors contribute to the identification of the working range of the sensor. Chapter 5 presents a calibration of the device and consequent assessment of reproducibility and repeatability of measurements. Calibration has dealt with every component singularly through specific measurements, such as measurements of the lateral profile of a rotating disc with cogs of different sizes, calibration of the accelerometer with a B&K calibrator and an oscillating bar, and the measurement of the profile of a surface with known characteristics.
Repeatability of measurements has been thoroughly verified, via repetitions of measurements of the same road profile; reproducibility has been checked by measuring the same road profile with another profilometer already available on the market, and owned by the Department of Civil Engineering of Università di Pisa.
Finally, an evaluation of the correlation between rolling noise and road texture is presented in chapter 6. Measurements of road texture provided by the profilometer are correlated with rolling noise emission on six different kinds of pavements: again, the chapter can be divided in two main sections. The former deals with the calculation of the correlation coefficient between parameters related to road texture, such as third-octave bands texture spectrum and rolling noise measured by CPX method, while the latter deals with the comparison of two existing models and the elaboration of a simple model to obtain the mean CPX broadband level of a road surface based the knowledge of its texture spectrum. The model uses two bands of the octave band texture spectrum, to predict rolling noise: the octave band centred at 5 mm, related to air pumping, is negatively correlated to sound generation, while the band at 80 mm shows a positive correlation, related to tyre vibrations. One of the main innovations of this model, despite the limited number of samples, is the usage of enveloped texture, which distinguishes road surfaces with sharp edges and deep grooves.
The correlations found confirm the theoretical knowledge of tyre/road noise, but also underline that a complete description of rolling noise must take into account not only road texture, but also the tyre itself. In fact, a resonance in the CPX noise spectrum at 80 km/h was also successfully identified.
The further perfectioning of the profilometer will permit deeper studies of the correlation between properties of road texture and rolling noise, and to develop a comprehensive model of tyre/road interaction noise: possible improvements are the simultaneous usage of two or more displacement sensors that profile adjacent lines, in order to reduce uncertainty in the evaluation of texture spectrum and the substitution of the current sensor with one that has a smaller spot size.