sensors Article

A Multisensing Setup for the Intelligent Tire Monitoring Francesco Coppo 1 , Gianluca Pepe 1 , Nicola Roveri 1,2 and Antonio Carcaterra 1, * 1 2

*

Department of Mechanical and Aerospace Engineering, Sapienza University of Rome, 00184 Rome, Italy; [email protected] (F.C.); [email protected] (G.P.); [email protected] (N.R.) CNIT, Consorzio Nazionale Interuniversitario per le Telecomunicazioni, 43124 Parma, Italy Correspondence: [email protected]; Tel.: +39-06-4458-5219; Fax: +39-06-484-854

Academic Editor: Vittorio M.N. Passaro Received: 10 January 2017; Accepted: 1 March 2017; Published: 12 March 2017

Abstract: The present paper offers the chance to experimentally measure, for the first time, the internal tire strain by optical fiber sensors during the tire rolling in real operating conditions. The phenomena that take place during the tire rolling are in fact far from being completely understood. Despite several models available in the technical literature, there is not a correspondently large set of experimental observations. The paper includes the detailed description of the new multi-sensing technology for an ongoing vehicle measurement, which the research group has developed in the context of the project OPTYRE. The experimental apparatus is mainly based on the use of optical fibers with embedded Fiber Bragg Gratings sensors for the acquisition of the circumferential tire strain. Other sensors are also installed on the tire, such as a phonic wheel, a uniaxial accelerometer, and a dynamic temperature sensor. The acquired information is used as input variables in dedicated algorithms that allow the identification of key parameters, such as the dynamic contact patch, instantaneous dissipation and instantaneous grip. The OPTYRE project brings a contribution into the field of experimental grip monitoring of wheeled vehicles, with implications both on passive and active safety characteristics of cars and motorbikes. Keywords: intelligent tire; smart tire; tire grip; autonomous vehicle; car monitoring

1. Introduction and State of the Art of “Smart Tires” Vehicle safety has been one of the subject of great interest for research centers, safety boards and car makers for many years. At the beginning of the 1980s, first electronic control systems that involved the contact between the tire and the road were introduced; one of them was the ABS, Antilock Braking System (ABS). This active control system and the ones that followed were based on an indirect estimation of the vehicle dynamics variables, such as forces, friction and load transfer to evaluate the tire-road friction factor by minimizing skid. In the last twenty years, research has moved towards a different approach [1–12], which is based on using newly conceived sensors [9,10] to pursue the measurement of some key variables involving the tire-road contact interface [13–16], such as pressure, temperature, contact-patch dimensions, or even better wheel load [17], acceleration, friction factor [18–20] or deformation [21,22]. These new hardware and related algorithms are capable of measuring new quantities and extract additional information with potential use for the enhancement of the performance of the vehicle control systems, also putting the basis for breakthrough in new control devices. All these systems rely on new sensors embedded in the tire to form the so so-called intelligent or smart tire [23–29], a system capable of monitoring directly and continuously these key parameters. The measured data are transferred to a logic control unit to be processed. Mathematical models able to put a light on the tire-road interface phenomena are also of great importance in this context. Sensors 2017, 17, 576; doi:10.3390/s17030576

www.mdpi.com/journal/sensors

Sensors 2017, 17, 576

2 of 23

The acquired data can be used to start an alert, or for monitoring and controlling the entire vehicle dynamics. In the last years, several different types of monitoring systems have been produced. For example, the TPMS (Tire Pressure Monitoring System), is the archetypal system of intelligent tire and one of the few technologies that has found commercial use. Many other systems, such as those based on strain measurements, or accelerations via optical or magnetic devices, show enormous potentialities, even if their industrial use meets some cost difficulties when involving large mass production. The key elements that can contribute to the success of a new technology in this field are related to: (i) a relative simplicity in the construction process, (ii) the availability of low invasive (small size) sensors, (iii) the chance of simple and effective sensors power supply, (iv) the need of a robust system for data transmission, and (v) the need of robust interpretative models for the acquired data. Among the solutions that obtained a good visibility in the context of intelligent tires, two macro-categories are worth mentioning, the former related to a direct acceleration measurement [6,8,13], the latter related to strain measurements [21,25–27], both of inner tire points. Works of Audisio [30], Morinaga [31] and Hannah [14] are more focused on a widespread industrial application of smart tire and employ commercial sensors, while others [8,13,15,25,32] use prototypal devices to study the nature of the phenomena involving the tire rolling. The Optyre technology [33–36], developed at the Department of Mechanical and Aerospace Engineering, Sapienza, University of Rome, has the focus to build-up a new smart tire for the real-time identification of the tire stress, of the residual grip and of the rolling resistance, mainly thanks to direct optical strain measurement. The present technology is based on the results of a new patent [36] and presents the following characteristics: (i) use of Fiber Bragg Grating-FBG sensors based strain measurements, with the twofold advantage: (ii) the energy supply to the sensor is conveyed directly via the light beam (that is no battery or electrical wires are needed), and (iii) the information is transmitted on the same light beam, intrinsically a wireless transmission (no need of any radio transmitter or receiver). Optyre is focused on the direct strain measurement, leading to some important advantages with respect to the acceleration based technologies. This choice is due to the difficulty in evaluating strain from an acceleration measurement that is an ill-conditioned inverse problem [37], as extensively discussed in [34]. This paper describes the new frontiers of the Optyre project, related to a new experimental setup showing the ability of the system to acquire very good quality data on the rolling tire, identifying specific features of the signal. The whole project is divided in three different sections: The first section of the project, presented in this paper, describes the development of a measurement system that relies on optical sensing lines [35], for a reliable acquisition system for the continuous measurement of the tire strain. This optical sensing line is obtained installing an array of FBG sensors in the tire carcass, provided with a series of smart solutions for the signal transmission. Furthermore, the optical measurement chain has been integrated with a parallel acquisition system composed by suitable sensors to analyze the optical measurement data. The second section regards the development of new models for the identification of the tire-road contact conditions, among them: The Brush Rod Beam model [34] where the study is mainly focused on the analytical definition of the phenomena at the tire-road interface, and the Optyre model [33], which employs a simpler analytical model to provide qualitative information concerning the rolling tire also outside the contact patch. The last section consists in the use of the acquired data as an input for the control logic that governs the vehicle dynamics, which is the third aim of the project. The paper is organized as follows. In Section 2, a description of the measuring hardware involved in the Optyre project is provided. The experimental campaign, carried out to test the performances of the developed prototype and to validate the theoretical models, is described in Sections 3 and 4. An overview of the technological applications of the Optyre project is presented in Section 5. Among all the existing technologies in this field, some relevant solutions, providing state of the art tools for

Sensors 2017, 17, 576 Sensors 2017, 17, 576

3 of 23 3 of 23

analyzing the contact patch phenomena in a rolling tire, are considered and compared to Optyre in analyzing the contact patch phenomena in a rolling tire, are considered and compared to Optyre in Section 6. Some conclusive remarks with future developments are outlined in Section 7. Section 6. Some conclusive remarks with future developments are outlined in Section 7. 2.2.Hardware HardwareSetup Setup The et al. al. [33], [33], is is aa multi-sensing multi-sensingdevice devicethat thatconsists consistsinina TheOptyre Optyre system, system, developed developed by by Carcaterra Carcaterra et atire tirewith withembedded embeddedFBG FBGand andother othersensors sensorsfor forthe thereal-time real-timeidentification identification of key parameters, such of key parameters, such as as the residual grip and the rolling resistance [35]. the residual grip and the rolling resistance [35]. 2.1. 2.1.The TheOptical OpticalSensing SensingLine Line The Theaim aimof ofthe theproject projectisisto todevelop developan anexperimental experimentalsystem systemable ableto tomeasure measurethe thecircumferential circumferential strain of a rolling tire, embedding FBG sensors within the tire carcass. strain of a rolling tire, embedding FBG sensors within the tire carcass. As Asaaquick quickreminder, reminder,Fiber FiberBragg BraggGrating Grating(FBG) (FBG)based basedsensors sensorsare aremade maderecreating recreatingaaspatially spatially recurring alternation of the glass refraction index in a single mode optical fiber. The grating turns the recurring alternation of the glass refraction index in a single mode optical fiber. The grating turns fiber into into an optical filter: in fact, a specifically narrowband infrared light fired the fiber an optical filter: in fact, a specifically narrowband infrared light firedinto intothe thefiber fiberisis reflected reflectedto toaaspectrum spectrumanalyzer, analyzer, while while the the rest rest of of the light is fully transmitted. Observing Observingthe thereflected reflectedspectrum, spectrum,ititisispossible possibleto torecognize recognizeaapeak peakvalue, value,called calledBragg Braggreflection reflection wavelength wavelength( (λBB),),as asshown shownin inFigure Figure1.1.

Figure1.1.Light Lightreflection reflectiondiagram diagramininaaFiber FiberBragg BraggGrating-FBG Grating-FBGsensor. sensor. Figure

The the Bragg Bragg reflection reflectionwavelength wavelengthdepends dependsonon the effective refractive index, , The value value of of the the effective refractive index, ne f f𝑛, 𝑒𝑓𝑓 and and on the grating period, , according to Equation (1), as presented by Antunes [38]: on the grating period, Λ, according to Equation (1), as presented by Antunes [38]: (1) λB = 2neff Λ λB = 2neff Λ (1) when the FBG sensor is stretched, or compressed, 𝜆𝐵 changes to 𝜆 , therefore, if the sensor is when the FBG is stretched, or compressed, λ, therefore, if the sensor is structure, integrated integrated withsensor the point to measure, for exampleλwelding it to the monitored B changesortopasting with the point to measure, for example welding or pasting it totothe structure,as the reflected the reflected wavelength changes are related to the strain and themonitored thermal variation, described wavelength changes are related to the strain and to the thermal variation, as described in Equation (2): in Equation (2): ∆λ ∆λ= (1 − ρ )ε + (α + η)ΔT λλB = (1 −eρe )ε + (α + η )∆T

(2)(2)

B

where 𝛥𝜆 = 𝜆 − 𝜆 , 𝜌𝑒 represents the photo-elastic coefficient, 𝜀 is the mechanical strain, 𝛼 is where ∆λ = λ − λ B , 𝐵ρe represents the photo-elastic coefficient, ε is the mechanical strain, α is the the thermal expansion of the silica, 𝜂 represents the temperature dependence of the refractive index, thermal expansion of the silica, η represents the temperature dependence of the refractive index, and and 𝛥𝑇 is the temperature variation. ∆T is the temperature variation. The FBG sensor is embedded within the tire: the cladding of the naked fiber is glued along the The FBG sensor is embedded within the tire: the cladding of the naked fiber is glued along inner side of the carcass, as in Figure 2. The selected adhesive is an epoxy glue with primer for the inner side of the carcass, as in Figure 2. The selected adhesive is an epoxy glue with primer for polyethylene and mostly polyolefins and silicone materials: Loctite®®406™ Prism Instant Adhesive polyethylene and mostly polyolefins and silicone materials: Loctite 406™ Prism Instant Adhesive and Loctite®®SF 770. and Loctite SF 770.

Sensors 2017, 17, 576 Sensors 2017, 17, 576

4 of 23 4 of 23

Sensors 2017, 17, 576

4 of 23

FBG Sensor FBG Sensor

Figure Figure2.2.FBG FBGsensor sensorglued glued on on the the carcass carcass of the tire. tire Figure 2. FBG sensor glued on the carcass of the tire Theinterrogator interrogatorisisresponsible responsible forboth both light light source source and and spectrum spectrum analyzer: The for analyzer: aa tunable tunablelaser laserfires fires selectedinfrared infraredwavelength wavelength light light into the the fiber fiber array. array. An An optical optical circulator is placed between the a aselected placedlaser between The interrogator is responsible into for both light source and spectrumcirculator analyzer: aistunable fires the FBGa sensor and the light source, which separates the fired beam from the reflected one, being the selected wavelength light into the fiber optical circulator is placed thelatter FBG sensor andinfrared the light source, which separates thearray. firedAn beam from the reflected one,between being the latter directed to the Optical Spectrum Analyzer (OSA). FBGto sensor and theSpectrum light source, which separates directed the Optical Analyzer (OSA). the fired beam from the reflected one, being the TM SmartScan [39] is the interrogator chosen for this project: the device can have The SmartFibers latter directed to the Optical Spectrum Analyzer (OSA). chosen for this project: the device can have TM The SmartFibers SmartScan [39] is the interrogator a multiplexed array upTMtoSmartScan 16 FBGs [39] per is channel, with upchosen to four channels, in have a 40 nm The SmartFibers the interrogator forindependent this project: the device can a multiplexed array up to 16 FBGs per channel, with up to four independent channels, in a 40 nm wide a multiplexed up to frequency 16 FBGs perischannel, up to four independent channels, innm), a 40 nm wide bandwidth.array The scan 2.5 kHzwith in full bandwidth mode (1528–1568 25 kHz bandwidth. The scan frequency is 2.5 kHz in full bandwidth mode (1528–1568 nm), 25 kHz reducing wide bandwidth. The scan frequency is 2.5 kHz in full bandwidth mode (1528–1568 nm), 25 kHz reducing the scan bandwidth. The case dimensions are 140 mm × 110 mm × 70 mm and the weight is the scan bandwidth. case dimensions are 140 mm × 110 mm × 70 mm andand the weight is is0.9 kg. reducing the scan The bandwidth. Thefrom case dimensions are 140 × 110 × 70 mm weight 0.9 kg. The information collected the interrogator ismm sent via mm Ethernet to the the workstation, on The information collected from the from interrogator is sent via Ethernet to thetoworkstation, on which a 0.9 akg. The information theallows interrogator is sent viathe Ethernet the workstation, on which suitable software iscollected installed. This a remote-lab as interrogator and the measuring suitable software is installed. allows remote-lab as the as interrogator and and the the measuring device which a suitable software is This installed. Thisa allows a remote-lab the interrogator measuring device can be far away from the workstation. can be far away the workstation. device can befrom far away from the workstation. The SmartSoft software is the frontend interface, where the user customizes the measurement The SmartSoft software thefrontend frontendinterface, interface, where customizes the the measurement The SmartSoft software is is the wherethe theuser user customizes measurement parameters, i.e., number of sensors, sample rate, data averaging, gain application, distance parameters, i.e., number of sensors, sample rate, data averaging, gain application, distance parameters, i.e., number of sensors, sample rate, data averaging, gain application, distance compensation, etc., and permits data log. compensation, permits datalog. log. compensation, etc.,etc., andand permits data A sketch of the optical inFigure Figure3.3. A sketch of the opticalsensing sensingchain chain is is provided provided in A sketch of the optical sensing chain is provided in Figure 3.

Figure logicscheme. scheme. Figure3.3.Interrogator Interrogator logic Figure 3. Interrogator logic scheme.

Sensors 2017, 17, 576

5 of 23

FBG sensors allow distributed sensing over significant area by multiplexing up to 16 sensors within each fiber and up to 64 sensors total. In the prototype system presented in this paper, only one optical line with a single FBG sensor equips the measurement device, whose length is 10 mm and Sensors 2017, 17, 576 5 of 23 λb = 1543 nm. This choice is also motivated by the fact that the installation of a single sensor avoids the spectra FBG sensors allow distributed sensing over significant area by multiplexing up to 16 sensors shouldering, which may arise when two or more sensors are multiplexed and large deformations are within each fiber and up to 64 sensors total. In the prototype system presented in this paper, only one involved. Given the low elastic modulus of the rubber in comparison with the silica, the bandwidth optical line with a single FBG sensor equips the measurement device, whose length is 10 mm and available for each FBG sensor needs to be as large as possible. b = 1543 nm. Namely, let us consider two FBG sensors on the same optical line, whose bandwidths are This choice is also motivated by the fact that the installation of a single sensor avoids the spectra contiguous, as in Figure 4. When the sensors are stretched, the original peaks in their respective shouldering, which may arise when two or more sensors are multiplexed and large deformations are spectrum (solid lines) are displaced (dotted lines), causing a potential overlap of the two spectra. involved. Given the low elastic modulus of the rubber in comparison with the silica, the bandwidth In this case, the wavelength analyzer is no longer capable to separate the two single contributions, available for each FBG sensor needs to be as large as possible. producing data loss from both sensors in the worst case.

Figure 4. FBG sensors overlapping phenomenon overview. Figure 4. FBG sensors overlapping phenomenon overview.

The maximum by the interrogator foroptical a sensor with λ B = bandwidths 1543 nm, in pure Namely, let usmeasurement consider twoallowed FBG sensors on the same line, whose are mechanical strain, is obtained fromthe Equation (3)are as follows: contiguous, as in Figure 4. When sensors stretched, the original peaks in their respective spectrum (solid lines) are displaced (dotted lines), causing a potential overlap of the two spectra. In ∆λmin/max this case, the wavelength analyzer isε min/max no longer to separate the two single contributions, = Kcapable (3) ε λ producing data loss from both sensors in the worst case. B The measurement allowed(about by the1.27 interrogator a sensor withand 𝜆𝐵 ∆λ = 1543is nm, in where Kε maximum is a constant of proportionality × 106 µε),for ∆λmin is 15 nm 25 nm. max pure mechanical strain, is obtained Equation (3) as follows: The calculated measurement range isfrom [–7400 µε, +20577 µε]. The stress limit of the optical fiber is about

Δ𝜆min/max 0.7 GPa, which, with a Young’s modulus of 16.56 GPa [40–42] produces a maximum strain range of (3) 𝜀min/max = 𝐾ε about 42300 µε to guarantee the fiber integrity. This range 𝜆B lies within the limits of the interrogator. where 𝐾𝜀 for is athe constant proportionality (about 1.27 × 106 𝜇𝜀), 𝛥𝜆 𝑚𝑖𝑛 is 15 nm and 𝛥𝜆 𝑚𝑎𝑥 is 25 2.2. The Device Opticalof Measurement nm. The calculated measurement range is [–7400 𝜇𝜀, +20577 𝜇𝜀]. The stress limit of the optical fiber Two0.7 main problems are involved in the realization of the present smart tire device: (i) the need to is about GPa, which, with a Young’s modulus of 16.56 GPa [40–42] produces a maximum strain guarantee the pressure sealtoavoiding anythe air leakage, still permitting the optical fiber the to come outoffrom range of about 42300 𝜇𝜀 guarantee fiber integrity. This range lies within limits the the inner of the tire, and (ii) the need of decoupling the circular motion of the FBG line within the tire, interrogator. from the part of the optical line connected to the interrogator, which is indeed integrated to the car body. general of the experimental device approaching the solution to problems (i) and (ii) 2.2. TheThe device for theview Optical Measurement is graphically presented in Figure 5. One can identify: the body of the valve, the support case of the Two main problems are involved the realization present tire device: (i) the need optical joint, and the kinematic linkagesinneeded to carry of thethe optical linesmart on board of the car. to guarantee pressure seal any air leakage, still permitting the this optical fiber tooutside come out The FBGthe sensor within theavoiding carcass is connected to the optical line, and is carried the from the inner of the tire, and (ii) the need of decoupling the circular motion of the FBG line within tire by a hole drilled into the rim. A specific feedthrough device has been designed, similar to a sealing the tire,which from is the part of the connected the interrogator, which is indeed integrated to valve, composed byoptical a seriesline of nuts drilledtoalong their symmetry axis, and mounted on the the body. The general rimcar as shown in Figure 6. view of the experimental device approaching the solution to problems (i) and (ii) is graphically presented in Figure 5. One can identify: the body of the valve, the support case After the insertion of the fiber, the internal channel of the feedthrough valve is sealed with silicon of optical andfiber the kinematic linkages needed to carry the optical line on board of the car. to the prevent thejoint, optical from abrupt bending. The presence of this polymer also inhibits possible The FBG sensor within the carcass is connected to the optical line, and this is carried outside the air leakages. tire by a hole drilled into the rim. A specific feedthrough device has been designed, similar to a sealing valve, which is composed by a series of nuts drilled along their symmetry axis, and mounted on the rim as shown in Figure 6. After the insertion of the fiber, the internal channel of the feedthrough valve is sealed with silicon to prevent the optical fiber from abrupt bending. The presence of this polymer also inhibits possible air leakages.

Sensors 2017, 17, 576

6 of 23

Sensors 2017, 17,17, 576 Sensors 2017, 576

of 23 6 of623

Decoupling

Kinematic link

device of FBG Decoupling fiber device of FBG

Kinematic link

fiber FBG valve FBG valve

FBG rotary FBG coupler rotary coupler Figure 5. Computer Aided Design-CAD of the wheel with in evidence the FBG valve, the FBG rotary Figure Computer Aidedlink. Design-CAD of the wheel with in evidence the FBG valve, the FBG rotary coupler5.and the kinematic Figureand 5. Computer Aidedlink. Design-CAD of the wheel with in evidence the FBG valve, the FBG rotary coupler the kinematic coupler and the kinematic link.

Standard FC/APC Standard connector FC/APC

Rim Rim

connector

FBG FBG terminal

sensor FBG

Valve FBG terminal (Nuts) Valve

Parts

sensor

Parts

(Nuts) OUTSIDE THE RIM OUTSIDE Figure 6. Exploded diagram of FBG valve. THE RIM Figure 6. Exploded diagram of FBG valve.

INSIDE THE RIM INSIDE THE RIM

An optical rotary coupler [43] was6.adopted coupleofthe rotating Figure Explodedtodiagram FBG valve. part of the fiber to the fixed one. This coupler was mounted within a purpose-designed specific support case. Figure 7 shows a picture of the joint. The trick consists in using a pair of lenses that transmit the light beam between the two

Sensors 2017, 17, 576

7 of 23

Sensors 2017, 17, 576

7 of 23

An optical rotary coupler [43] was adopted to couple the rotating part of the fiber to the fixed rotary coupler [43] was aadopted to couple the rotating part of the Figure fiber to7the fixeda one. An Thisoptical coupler was mounted within purpose-designed specific support case. shows one. This within purpose-designed specific supportthe case. Figure shows a picture of coupler the joint.was Themounted trick consists inausing a pair of lenses that transmit light beam7 between Sensors 2017, 17, 576 7 of 23 picture the joint.optical The trick consists inany using a pair of lenses transmit the two of separated lines, without mechanical contactthat between the the twolight fiberbeam ends. between The two the twoare separated lines, without between the two fiber ends. The of two lenses indeed optical each integrated withany twomechanical cylindricalcontact elements, respectively, that are parts a separated optical lines, without any mechanical contact between the two fiber ends. The two lenses lenses are indeed integrated with cylindrical elements, respectively, are parts of a mechanical journaleach bearing. Namely, onetwo of the cylinder is integrated with the carthat chassis, the other are each integrated with two cylindrical that are of2000 a mechanical mechanical journal bearing. Namely, oneturn of theatelements, cylinder isrespectively, integrated the parts car chassis, the other withindeed the wheel. This optical joint can a maximum angularwith speed about rpm. Its journal bearing. Namely, one of the cylinder is integrated with the car chassis, the other with the wheel. with the wheel. This optical joint can turn at a maximum angular speed about 2000 rpm. Its dimensions and weight are very small, with the maximum diameter about 1 cm, and the weight about This joint can turn atof a the maximum angular speed about 2000 rpm. Its 1dimensions weight are dimensions and weight are very small, the maximum diameter about cm, theand weight about 10 g.optical The optical terminals rotarywith coupler are connected to the optical lineand coming out from the very small, with the maximum diameter about 1 cm, and the weight about 10 g. The optical terminals 10 g.and The the optical terminals of the rotary are connected to the line coming out adopted, from the tire, optical line towards to the coupler interrogator, respectively. Inoptical the mounting scheme of the rotary coupler are connected to the optical line coming out from the tire, and the optical tire, andof the towards to be thealigned interrogator, respectively. thecross-section mounting scheme adopted, the axis theoptical opticalline coupler must with the wheel axis In (see in Figure 5).line towards to the interrogator, respectively. In the mounting scheme adopted, the axis of the optical the axis of the optical coupler must be aligned with the wheel axis (see cross-section in Figure 5). coupler must be aligned with the wheel axis (see cross-section in Figure 5).

Figure Figure 7. 7. Picture Picture of of the the optical optical rotary rotary coupler. coupler Figure 7. Picture of the optical rotary coupler

The third thethe device, represented in Figure 5, is a 3D kinematic chain linkage (articulate thirdelement elementof of device, represented in Figure 5, is a 3D kinematic chain linkage The third element of thethat device, represented intranslation Figure 5, of istranslation a 3D kinematic quadrilateral with ball joints) follows the vertical the wheel, while impeding the (articulate quadrilateral with ball joints) that follows the vertical of the chain wheel,linkage while (articulate quadrilateral ball joints) that the vertical translation of the steering wheel, while rotation of the second of optical joint. When mounted onWhen the steering fronton wheel, the kinematic impeding rotationlens ofwith thethe second lens of thefollows optical joint. mounted the front impeding the rotation of second lensfor of optical joint.axis. When on the steering front linkages also allow for thethe wheel about the vertical A mounted rendering and the final device wheel, the kinematic linkages alsorotation allow the wheel rotation about the vertical axis. A rendering wheel, linkages also the wheel rotation mounted onkinematic the car is shown in Figure and thethe final device mounted on theallow car8.isfor shown in Figure 8. about the vertical axis. A rendering and the final device mounted on the car is shown in Figure 8.

(a) (b) (c) (a) (b) (c) Figure perspective; (c) (c) actual actual prototype). prototype). Figure 8. 8. Support Support case case onboard onboard installation installation ((a) ((a) CAD CAD model; model; (b) (b) perspective; Figure 8. Support case onboard installation ((a) CAD model; (b) perspective; (c) actual prototype).

2.3. Phonic and Other Other Sensors Sensors 2.3. Phonic Wheel, Wheel, Accelerometer Accelerometer and 2.3. Phonic Wheel, Accelerometer and Other Sensors The Optyre employs other other sensors sensors besides besides FBG, FBG, which which are are connected to aa data The Optyre project project employs connected to data acquisition acquisition The Optyre project employs other sensors besides FBG, which are connected to a data hardware that is the LMS Scadas Mobile [44]. The measurement system is connected to host hardware that is the LMS Scadas Mobile [44]. The measurement system is connectedacquisition to aa host hardware that is the LMS Scadas Mobile [44]. The measurement system is connected to a host workstation through through an an Ethernet Ethernet connection. connection. workstation workstation through an Ethernet connection. The additional sensors chain is composed by: (i) a magnetic Hall effect sensor for the angle and angular speed detection, (ii) a uniaxial accelerometer for the measurement of the vertical acceleration of the wheel axis, (iii) a dynamic temperature sensor for the detection of the thread temperature, and

Sensors 2017, 17, 576

8 of 23

The additional sensors chain is composed by: (i) a magnetic Hall effect sensor for the angle and angular speed detection, (ii) a uniaxial accelerometer for the measurement of the vertical acceleration of the wheel axis, (iii) a dynamic temperature sensor for the detection of the thread temperature, and Sensors 2017, 17, 576 8 of 23 (iv) a GPS system for the global position detection and for the timing reference of the vehicle on which the hardware is installed. few sensors, such the tachymetric loggingand system and the accelerometer, linked the (iv)A a GPS system for theas global position detection for the timing reference of theare vehicle ontowhich support case, asisshown in Figure 9. the hardware installed. The tachymetric measurement system is composed by a phonic wheel and a magnetic Hall effect A few sensors, such as the tachymetric logging system and the accelerometer, are linked to the sensor. The phonic wheelinisFigure obtained support case, as shown 9. from a galvanized ferritic steel disk with an alternation of 100 vacuum and 100 bulk sections for the whole turn, recreating a 1.8° spatial grating, as in Figure 9.

(a)

(b)

(c)

Figure 9. Phonic wheel design: in (a) the perspectives; in (b) the assembly; and (c) a picture of the Figure 9. Phonic wheel design: in (a) the perspectives; in (b) the assembly; and (c) a picture of the phonic wheel installed on the support case. phonic wheel installed on the support case.

TheHall tachymetric measurement system is composed a phonic wheel and a magnetic The effect sensor selected for the project is theby sensor model AA020020 [45] by Hall Elen effect srl, sensor. phonic wheel is obtained from of a galvanized ferritic steel disk with an alternation which hasThe a maximum acquisition frequency 15 kHz, corresponding to a maximum detectableof 100 vacuum and 100 bulk for the whole turn, recreating 1.8◦considering spatial grating, as in Figure speed of 148 km/h, given thesections actual dimension of the phonic wheelaand at least three data9. The sensor selected for the project is the sensor model AA020020 [45] by Elen srl, which samples inHall eacheffect vacuum or bulk section. hasThe a maximum acquisition ofwheel 15 kHz, to support a maximum speed Hall effect sensor andfrequency the phonic arecorresponding mounted on the case,detectable on the stator andof 148 km/h, given the actual dimension of the phonic wheel and considering at least three data samples on the rotor, respectively, as shown in Figure 9. in each or bulk The vacuum accelerometer is section. a PCB® axial ICP® accelerometer (Integrated Circuit Piezoelectric 352C33) The Hall effect sensor the[0.5 phonic are measurement mounted on the support onsensitivity the stator and with acquisition within the and range Hz, wheel 10 kHz], range ±50 case, g and of onmV/g. the rotor, as shown in Figure 100 Thisrespectively, sensor is installed on the stator 9. side of the case, to detect the z-axis acceleration of the ® ® wheel.The accelerometer is a PCB axial ICP accelerometer (Integrated Circuit Piezoelectric 352C33) with acquisition within the range [0.5 10 kHz], measurement rangeto±50 g and sensitivity The data acquisition hardware is Hz, provided with a GPS system, record the historicalof 100 mV/g.ofThis installed on thedifferent stator side of the case, to detect theadoption z-axis acceleration movements the sensor car andisto synchronize acquisition systems, by the of a globalof the wheel. clock. The installed GPS records longitude, latitude, altitude, global and uniaxial speed and the The acquisition hardware is provided GPSbeen system, to record the historical movements number of data achieved satellites. The antenna of thewith GPSahas installed on-board. of the andtemperature to synchronize different acquisition bythe thethread adoption of a globalThe clock. The An car optical sensor has been installedsystems, to monitor temperature. sensor installed GPS records longitude, latitude, altitude, global and uniaxial speed and the number is the Optris CSmicro [46] LT-15 pyrometer, which has a measurement range between 0 and 350 °C,of achieved satellites. The the GPS has been installed on-board. and a response time of 30 antenna ms. Thisof sensor is connected to the Scadas. The sensitivity is 10 mV/°C. The temperature sensor hasand beenthe installed to monitor thread The sensor sensorAn hasoptical been installed on the vehicle, detection spot has the an area of temperature. 1 cm2 of the tire thread ◦ C, is the Optris CSmicro [46] LT-15 pyrometer, which has a measurement range between 0 and 350 in the proximity of the footprint side. The temperature is measured in the neighborhood of the ◦ C. and a response time of 30patch, ms. This sensor is connected to the Scadas. mV/to leading edge of the contact in order to detect the temperature whileThe the sensitivity FBG sensorisis10 closer 2 The sensorarea, has and beentoinstalled the vehicle, and the the detection an area of 1of cm the tire the contact preserveon sensor life: whenever sensor spot was has placed in front theoftrailing thread in the proximity of the footprint side. The temperature is measured in the neighborhood of the edge of the tire, it would be exposed to debris. leading edge oftire thewill contact patch, in order detect the temperature while sensor is closer The smart be equipped in thetonear future with a system [47]the forFBG the detection of theto the contact area, and to preserve sensor life: whenever the sensor was placed in front of the trailing drive-wheel torque, which measures the strain of the rotating semi-axle, and then, according to edge of the tire,it itprovides would be exposed to debris. Hoffmann [48], the measurement of the torque. The measure is carried out with the use The smart tiremodule will be to equipped in an theeasy neardata future with a system for chain the detection of the of a radio-frequency guarantee transmission. This[47] supply is composed drive-wheel torque, which measures the strain of the rotating semi-axle, and then, according to Hoffmann [48], it provides the measurement of the torque. The measure is carried out with the use of a radio-frequency module to guarantee an easy data transmission. This supply chain is composed of two main parts, one installed on the axle, and one on the chassis, as in Figure 10. The output of this system is the output voltage of the Wheatstone bridge on which the strain gauges are connected.

Sensors2017, 2017,17, 17,576 576 Sensors

of23 23 99of

oftwo two2017, main parts, one installed on the axle, and one on the chassis, as in Figure 10. The output of this of main this Sensors 17,parts, 576 one installed on the axle, and one on the chassis, as in Figure 10. The output of 9 of 23 systemisis the theoutput output voltage voltage of of the theWheatstone Wheatstone bridge bridgeon on which which the thestrain straingauges gaugesare are connected. connected. system

Figure10. 10.Radio RadioFrequency Frequencydata datatransmission transmissionsystem, system,from fromPCB PCB®®® datasheet datasheet[47]. [47]. Figure

An ICP®®® sensor sensor for detection detection of the the dynamic pressure pressure will be be installed in in the inner inner tube and and An An ICP ICP sensor for for detection of of the dynamic dynamic pressure will will be installed installed in the the inner tube tube and connected to a parallel branch of the radio-frequency system. The same supply chain described for connected to aa parallel parallelbranch branchofofthe theradio-frequency radio-frequency system. The same supply chain described for connected to system. The same supply chain described for the the torque torque sensor sensor has has been been chosen chosen for for the the data data information information transmission transmission regarding regarding the the dynamic the torque sensor has been chosen for the data information transmission regarding the dynamic dynamic pressure pressure on the internal chamber of the tire. pressure on the internal chamber of the tire. on the internal chamber of the tire. Finally, Figure Figure 11 11shows showsan anoverview overview of of the the hardware hardwaresetup setup of ofthe the project. project. Finally, Finally, Figure 11 shows an overview of the hardware setup of the project.

Figure 11. Sensing line conceptual scheme. Figure11. 11.Sensing Sensingline lineconceptual conceptualscheme. scheme. Figure

Sensors 2017, 17, 576 Sensors 2017, 17, 576

10 of 23 10 of 23

The SCADAS data acquisition system can be configured in different ways, and the version used for The the SCADAS project isdata equipped with system 16 VB8can ports for the connection with theand sensors. In terms acquisition be configured in different ways, the version usedof maximum frequency, theports hardware detecting up sensors. to 102 kHz, splittable in three for the projectacquisition is equipped with 16 VB8 for theallows connection with the In terms of maximum different ranges, i.e., Static, Vibrational, and Acoustic, according to the nature of the inspected acquisition frequency, the hardware allows detecting up to 102 kHz, splittable in three different ranges, phenomenon, while the refreshes the datatoatthe a frequency about 4 Hz and filtered at 1while Hz. the i.e., Static, Vibrational, andGPS Acoustic, according nature of the inspected phenomenon, Actually,the the Optyre project provides solution GPS refreshes data at a frequency about 4an Hzalternative and filtered at 1 Hz.to miniaturize the hardware. The authors are designing a single compact device containing all the to measurement all in one.The The Actually, the Optyre project provides an alternative solution miniaturizechain, the hardware. new device will exchange data through wireless connections it will be supplied authors are designing a single compact device containing all theand measurement chain, alldirectly in one.inside The the tire by a solution based on the study of the waveguides generated in the tire [49], in order new device will exchange data through wireless connections and it will be supplied directly inside theto realize micro-scale devices exploit the energy pumping phenomenon, as presented inrealize [50], to tire by a solution based on thethat study of the waveguides generated in the tire [49], in order to harvest and collectthat energy from and wave-propagation micro-scale devices exploit thevibrations energy pumping phenomenon, asphenomena. presented in [50], to harvest and collect energy from vibrations and wave-propagation phenomena. 3. Tuning of the Acquisition Hardware 3. Tuning of the Acquisition Hardware The experimental campaign begins with the validation of the measurement chain. The phonic wheel somecampaign irregularities due to the the validation tolerancesofofthe manufacturing andphonic to the Thepresents experimental begins with measurement process chain. The assembly, causing not negligibledue error in the angle span of the single element, a calibration wheel presents someairregularities to the tolerances of manufacturing processtherefore and to the assembly, of the tachymeter is required. causing a not negligible error in the angle span of the single element, therefore a calibration of the A correction algorithm has been built up to identify the phase shift between bulks and voids. In tachymeter is required. the A matter in question, the has tachymetric beenthe putphase on a shift constant speedbulks rotating correction algorithm been builtsystem up to has identify between and system, voids. morphology of every fullsystem or empty aim ofspeed calculating the correct Inidentifying the matter the in question, the tachymetric haselement, been putwith on a the constant rotating system, phase shiftthe towards a least squares Figure 12 shows thethe raw voltage output from the Hall identifying morphology of everytechnique. full or empty element, with aim of calculating the correct effectshift sensor. Starting from the data, the algorithm identifies theraw timevoltage instantoutput in which there is the phase towards a least squares technique. Figure 12 shows the from the Hall passage between bulkfrom and the voiddata, or vice as shown in Figure 13. The angular distribution the effect sensor. Starting theversa, algorithm identifies the time instant in which there isofthe elements is plotted forand a single in Figure 14,shown where in theFigure angular of the elements of swings passage between bulk void turn or vice versa, as 13.dimension The angular distribution the betweenis1.7° and for 1.9° with an average value 1.8°.the After the statistical evaluation of the swings angular elements plotted a single turn in Figure 14, of where angular dimension of the elements ◦ with ◦ . After the angular dimensions the1.9 single element, it is possible correction of each The between 1.7◦ of and an average value of to 1.8evaluate statistical evaluation of thesector. angular results of the correction algorithm are shown to in evaluate Figure 15, where it iscorrection possible toofnotice the benefits dimensions of the single element, it is possible the angular each sector. The of theof correction algorithm on theare measure angular speed. results the correction algorithm shown of in the Figure 15, where it is possible to notice the benefits of the correction algorithm on the measure of the angular speed.

Figure Figure12. 12.Sample Sampleofofraw rawdata datafrom fromthe theHall Halleffect effectsensor. sensor.

Sensors 2017, Sensors 2017, 17, 17, 576 576 Sensors Sensors 2017, 2017, 17, 17, 576 576

11 of 23

11 23 of 23 11 11 of of 23

Figure 13. Bulk/void Figure13. 13.Bulk/void Bulk/void shift shift identification. Figure 13. Bulk/void shift identification. Figure shiftidentification. identification.

Figure angulardispersion. dispersion. Figure 14. Bulk/void Figure14. 14.Bulk/void Bulk/void angular angular dispersion.

Figure 14. Bulk/void angular dispersion.

Figure 15. Angular speed correction algorithm results: angular speed vs correct angular speed. Figure Angularspeed speedcorrection correctionalgorithm algorithm results: results: raw raw speed. Figure 15.15. Angular raw angular angularspeed speedvsvscorrect correctangular angular speed.

Figure 15. Angular speed correction algorithm results: raw angular speed vs correct angular speed.

Sensors 2017, 17, 576 Sensors 2017, 17, 576

12 of 23 12 of 23

It should noted output of the interrogator cannot connected as an input to the It should be be noted thatthat thethe output of the interrogator cannot be be connected as an input to the SCADAS system because of the digital nature of the data transmission system interrogator. SCADAS system because of the digital nature of the data transmission system of of thethe interrogator. Therefore, to obtain comparable data from different acquisition hardware, devices need Therefore, to obtain comparable data from thethe different acquisition hardware, thethe twotwo devices need to synchronized. be synchronized. This task was manually carried operator at the stage of the project, to be This task was manually carried outout byby thethe operator at the stage of the project, causing a small time delay between of data logging: caused reaction time causing a small time delay between thethe twotwo setssets of data logging: oneone caused by by thethe reaction time of of operators in charge fordata therecording, data recording, thecause other cause is the delay the two different thethe operators in charge for the the other is the delay of the two of different software software on starting the data logging. The proposed is the stimulation of thewith system packages onpackages starting the data logging. The proposed solution issolution the stimulation of the system with a recognizable perturbation, such as a hammer hit on the tire with the tire wheel at rest. a recognizable perturbation, such as a hammer hit on the tire with the tire wheel at rest. The timeThe time histories of the sets arein shown Figure 16,the where the redshow circles time instants histories of the two datatwo setsdata are shown Figurein16, where red circles theshow timethe instants after afterthe which the hammer hitplace takesin place in respect the origin the time A developed algorithm which hammer hit takes respect to theto origin of theoftime axes.axes. A developed algorithm provides proper shift of the time axes, so that hammer hitthe is the for both axes. provides thethe proper shift of the time axes, so that the the hammer hit is zerozero timetime for both timetime axes.

Figure Raw data appearance before synchronization: (top) time history of the vertical Figure 16. 16. Raw data appearance before synchronization: (top) thethe time history of the vertical acceleration, in g, wheelhub, hub,and and(bottom) (bottom)the thetime timehistory history of of the the peak acceleration, in g, of of thethe wheel peak wavelength wavelength 𝜆λ𝐵B of of the the FBG FBG sensor, sensor, in in nm. nm.

4. Experimental Campaign 4. Experimental Campaign described measurement device permits to observe time which elastic TheThe described measurement device permits to observe for for thethe firstfirst time which areare thethe realreal elastic phenomena place inside in real operating rolling conditions. phenomena thatthat taketake place inside thethe tiretire in real operating rolling conditions. All the previously described components of the measurement chain used to correlate All the previously described components of the measurement chain areare used to correlate thethe circumferential strain acquired by FBG sensors, directly to the tire temperature and pressure, wheel circumferential strain acquired by FBG sensors, directly to the tire temperature and pressure, wheel vertical displacement, angular position speed of the vertical displacement, angular position andand speed of the tire.tire. Two types of measurements characterize the experimental campaign: performed with Two types of measurements characterize the experimental campaign: oneone performed with thethe vehicle in steady state motion, and the other involves accelerated/decelerated motion: both vehicle in steady state motion, and the other involves accelerated/decelerated motion: both are carriedare along atrajectory rectilinear trajectory and to keepininto phaseonly of the outcarried along aout rectilinear and real roads, to real keeproads, into account, this account, phase of in thethis analysis, analysis, only forces and torques involved with the longitudinal motion of the car. forces and torques involved with the longitudinal motion of the car. Figure 17 shows time history of the wavelength changes acquired FBG sensor, during Figure 17 shows thethe time history of the wavelength changes acquired by by FBG sensor, during oneone single revolution of the tire. The y-axis is directly proportional to the circumferential strain of the tire, single revolution of the tire. The y-axis is directly proportional to the circumferential strain of the tire, as shown by Equation (2). As far as the sensor is roughly in correspondence of the center of the contact as shown by Equation (2). As far as the sensor is roughly in correspondence of the center of the contact patch, strain achieves maximum value, otherwise, when sensor is far from contact patch, thethe strain achieves thethe maximum value, otherwise, when thethe sensor is far from thethe contact patch the strain remains close to zero and the strain values are main influenced by temperature and inflation pressure.

Sensors 2017, 17, 576 Sensors 2017, 17, 576

13 of 23 13 of 23

Figure 18 shows the polar plot of the circumferential strain after several tire rotations in a steady 13 of 23 state motion: curves belonging to different rotations mostly coincide. Figure 19 summarizes how the patch the strain remains close to zero and the strain values are main influenced by temperature and measured tire strain is correlated to tire angular position, detected by the phonic wheel system. inflation pressure. Figure 18 shows the polar plot of the circumferential strain after several tire rotations in a steady state motion: curves belonging to different rotations mostly coincide. Figure 19 summarizes how the measured tire strain is correlated to tire angular position, detected by the phonic wheel system. Sensors 2017, 17, 576

Figure Figure17. 17.Reflected Reflectedwavelength wavelengthfor foraacomplete completewheel wheelturn. turn.

Figure 18 shows the polar plot of the circumferential strain after several tire rotations in a steady Figure 17. Reflected wavelength for a complete wheel turn. state motion: curves belonging to different rotations mostly coincide. Figure 19 summarizes how the measured tire strain is correlated to tire angular position, detected by the phonic wheel system.

Figure 18. Polar plot of circumferential strain (𝜇 𝑠𝑡𝑟𝑎𝑖𝑛) as function of the wheel rotation angle (𝜃). Figure 18. Polar plot of circumferential strain (µ strain) as function of the wheel rotation angle (θ ).

Figure 18. Polar plot of circumferential strain (𝜇 𝑠𝑡𝑟𝑎𝑖𝑛) as function of the wheel rotation angle (𝜃).

Sensors 2017, 2017, 17, 17, 576 576 Sensors Sensors 2017, 17, 576

14 14 of of 23 23 14 of 23

Plot of strain and FBG sensor position as function of the wheel rotation. Figure 19. Plot Figure 19. Plot of strain and FBG sensor position as function of the wheel rotation.

Figure 20 shows shows the the strain strain detected detected by by the the FBG FBG sensor sensor in in aa 10 10 ss steady steady state state motion motion at at 40 40 km/h, km/h, Figure 20 Figure 20 shows the strain motion, detectedand by the FBG sensor in a 10 s steady state of motion at21 40 taken km/h, Figure 21 shows an accelerated in Figure 22 shows a magnification Figure Figure 21 shows an accelerated motion, and in Figure 22 shows a magnification of Figure 21 taken Figure 21 shows an accelerated motion, and in Figure 22 shows a magnification of Figure 21 taken around 163.5and and182.5 182.5 s, which shows a brief stop followed by a resumption of theFigure motion. around 163.5 s, which shows a brief stop followed by a resumption of the motion. 22 around 163.5 and 182.5 s, which shows a brief stop followed by a resumption of the motion. Figure 22 shows of the tire of the vehicle, asbyindicated by thetrend decreasing shows the coolingthe of cooling the tire during the during stop of the the stop vehicle, as indicated the decreasing of the Figure 22 the cooling of the during the stopin ofEquation the vehicle, trend theshows signal, the tire second addendum (2).as indicated by the decreasing signal,ofaccording toaccording the secondtoaddendum in Equation (2). trend of the signal, according to the second addendum in Equation (2).

Figure 20. 20. Strain log for the vehicle vehicle in in steady steady motion motion at at 40 40 km/h. km/h. Figure Figure 20. Strain log for the vehicle in steady motion at 40 km/h.

17, 576 576 Sensors 2017, 17, Sensors 2017, 17, 576

15 of 23 15 of 23

Figure 21. FBG wavelength log for a random speed straight run. Figure Figure 21. 21. FBG FBG wavelength wavelength log log for for aa random random speed speed straight straight run. run.

Figure 22. Detail of Figure 21 between 168 s and 178 s with with aa stop-start stop-start sequence sequence between 172 s and Figure 22. Detail of Figure 21 between 168 s and 178 s with a stop-start sequence between 172 s and 174.5 s. s. 174.5 s.

Figure 23 23 shows aa five five seconds window window that summarizes summarizes the recorded recorded data from from all the the sensors Figure Figure 23 shows shows a five seconds seconds window that that summarizes the the recorded data data from all all the sensors sensors between 21.5 s and 26.5 s for the accelerated/decelerated motion. The data show an acceleration from between between 21.5 21.5 ss and and 26.5 26.5 ss for for the the accelerated/decelerated accelerated/decelerated motion. motion. The The data data show show an an acceleration acceleration from from 0 km/h to approximately 8.5 km/h, inter alia, the motion starts at about 22 s, where the encoder signal, 00 km/h to approximately 8.5 km/h, inter alia, the motion starts at about 22 s, where the encoder signal, km/h to approximately 8.5 km/h, inter alia, the motion starts at about 22 s, where the encoder signal, shown in in thesecond second rowofofthe thepicture, picture, switch square wave, caused by approaching the approaching the shown switch thethe square wave, caused by the of theof shown inthe the secondrow row of the picture, switch the square wave, caused by the approaching ofHall the Hall effect sensor from a vacuum to a bulk section. The deformation of the FBG sensor is represented effect sensor from a vacuum to a bulk section. The deformation of the FBG sensor is represented in the Hall effect sensor from a vacuum to a bulk section. The deformation of the FBG sensor is represented in the firstInrow. In the picture it is possible to recognize up passages to three passages of the in the first row. the picture is possible to recognize up to three of the sensor in sensor the footprint. in the first row. In the it picture it is possible to recognize up to three passages of the sensor in the footprint. The shows third row the angle increase to of the motion, while row the fourth The third row the shows angle due to the due start ofthe thestart motion, while the fourth showsrow the footprint. The third row showsincrease the angle increase due to the start of the motion, while the fourth row shows the angular speed of the wheel; from the second to the fourth row the data are collected from shows the angular speed of the wheel; from the second to the fourth row the data are collected from

Sensors 2017, 17, 576

16 of 23

Sensors 2017, 16 23 Sensors 2017,17, 17,576 576 wheel; from the second to the fourth row the data are collected from the 16ofofphonic 23 angular speed of the wheel system. The graph representing the temperature shows that at the beginning of the motion the the thephonic phonicwheel wheelsystem. system.The Thegraph graphrepresenting representingthe thetemperature temperatureshows showsthat thatatatthe thebeginning beginningof ofthe the temperature initially increases, after which the temperature stabilizes to a new value. The last row motion motionthe thetemperature temperatureinitially initiallyincreases, increases,after afterwhich whichthe thetemperature temperaturestabilizes stabilizesto toaanew newvalue. value.The The shows the vertical acceleration of the wheel can bethat directly associated with thewith vertical load acting last row shows the of the can directly the last row shows thevertical verticalacceleration acceleration ofthat thewheel wheel that canbe be directlyassociated associated with thevertical vertical on the wheel, and then deforms the contact patch. Furthermore, it is possible to follow the vehicle load loadacting actingon onthe thewheel, wheel,and andthen thendeforms deformsthe thecontact contactpatch. patch.Furthermore, Furthermore,ititisispossible possibleto tofollow follow usingthe a Global System. The Figure shows a24 picture the geo-referenced data, indata, terms of vehicle using Position System. The Figure aapicture of the vehiclePosition usingaaGlobal Global Position System.24 The Figure 24shows showsof picture ofthe thegeo-referenced geo-referenced data, in of plotted with the image in ItItisispossible to latitude and longitude, plotted with the satellite in the background. It is possible to identify interms terms oflatitude latitudeand andlongitude, longitude, plotted withimage thesatellite satellite image inthe thebackground. background. possible to the identify the area in the background of the picture: the Vehicle Dynamics and Mechatronics identify the area in the background of the picture: the Vehicle Dynamics and Mechatronics area in the background of picture: the Vehicle Dynamics and Mechatronics Laboratory of Sapienza, Laboratory of in located in Laboratory ofSapienza, Sapienza, University inRome, Rome, located inCisterna Cisterna diLatina, Latina,Latina, Latina,Italy. Italy. University in Rome, locatedUniversity in Cisterna di Latina, Latina, Italy. di

Figure 23. Multi-sensor data logs: Strain, Encoder output voltage, Angle, Rotational speed, Figure Figure 23. 23. Multi-sensor Multi-sensor data data logs: logs: Strain, Strain, Encoder Encoder output output voltage, voltage, Angle, Angle, Rotational Rotational speed, speed, Temperature, and Wheel hub vertical acceleration, respectively. Temperature, and Wheel hub vertical acceleration, respectively. Temperature, and Wheel hub vertical acceleration, respectively.

Figure Figure24. 24.GPS GPStracking trackingdata datalog. log.

Figure 24. GPS tracking data log.

Sensors 2017, 17, 576 Sensors 2017, 17, 576

17 of 23 17 of 23

5. Reconstruction Reconstruction of of Fundamental Fundamental Tire TirePhenomena Phenomena 5. The data data acquired acquired with with the the built-in built-in hardware hardware is is processed processed with with algorithms algorithms developed developed in in related related The works [33–35], for the reconstruction of the phenomena at the contact patch, such as the angular works [33–35], for the reconstruction of the phenomena at the contact patch, such as the angular position of of the the maximum maximum detected detected deformation, deformation, the the footprint footprint length, length, the the leading leading and and trailing trailing edges, edges, position the transition abscissa between stick and slip region and the evaluation of the rolling resistance. the transition abscissa between stick and slip region and the evaluation of the rolling resistance. An important important parameter parameter for for the the study study of of the the phenomena phenomena at at the the contact contact patch patch is is the the angular angular An positionθ 𝜃𝑚𝑎𝑥at at which maximum circumferential deformation within the tire is reached, as position which thethe maximum circumferential deformation within the tire is reached, as shown max shown in Figure 25, for a few rotations in constant speed motion. For each tire revolution, the peak in Figure 25, for a few rotations in constant speed motion. For each tire revolution, the peak of the of the circumferential deformation is detected the associated angularθposition 𝜃𝑚𝑎𝑥 𝑖 . circumferential deformation is detected together together with the with associated angular position max i . The polar The polar of the angular positioning of each detected 𝜃𝑚𝑎𝑥 𝑖 are reported in Figure 26. This plots of theplots angular positioning of each detected θmax i are reported in Figure 26. This distribution distribution of points permits to derive a probability density function-PDF 𝑝𝑑𝑓(𝜃 is 𝑚𝑎𝑥 ), which in of points permits to derive a probability density function-PDF pd f (θmax ), which is represented represented in Figure 25. Thethat PDF shows both and the average and the value maximum of lie thein curve Figure 25. The PDF shows both the that average the maximum of thevalue curve the lie in the neighborhood of 180 degrees, but introducing a clear displacement of the pdf peak in the neighborhood of 180 degrees, but introducing a clear displacement of the pdf peak in the direction direction of motion of the wheel center, making the distribution asymmetrical. This effect amounts of motion of the wheel center, making the distribution asymmetrical. This effect amounts to the to the unavoidable presence of dissipation effects that perturb static symmetric distribution of the unavoidable presence of dissipation effects that perturb the staticthe symmetric distribution of the vertical vertical load, producing the well known displacement of the vertical force application point towards load, producing the well known displacement of the vertical force application point towards the the direction ofwheel the wheel center direction of the center [51].[51].

Figure 25.25. PDF and polar plot of of strain peak angular position and plot of of local speed over time Figure PDF and polar plot strain peak angular position and plot local speed over time.

The same same statistical statistical approach to The approach also also suggests suggeststhe theestimation estimationofofthe thefootprint footprintlength. length.According According [33], the leading and the trailing edges, defining the initial and final points of the footprint, to [33], the leading and the trailing edges, defining the initial and final points of the footprint, correspond to to the thethe deformation: in analytical terms, the correspond the maxima maxima of ofthe thesecond secondtime timederivative derivativeofof deformation: in analytical terms, deformation speed rate exhibits stationary points at the two edges. The polar plot shows the positions the deformation speed rate exhibits stationary points at the two edges. The polar plot shows the of these two points (black for(black the leading green forgreen the trailing edge, respectively) in a set in of positions of these two points for the edge, leading edge, for the trailing edge, respectively) identicalidentical rotations. It is interesting to note that the two points aligned along two anominally set of nominally rotations. It is interesting to note thatsets theof two sets are of points are aligned radii that are not symmetrical with respect to the ideal tangent point of the wheel. In fact, the footprint along two radii that are not symmetrical with respect to the ideal tangent point of the wheel. In fact, appears to be,appears once again, in the directioninofthe thedirection wheel center a consequence the footprint to be,displaced once again, displaced of thespeed, wheelascenter speed, as of a the dissipation For eacheffects. set of points oneset canofbuild theone corresponding Figure 26 shows consequence of effects. the dissipation For each points can build thePDF. corresponding PDF. both the at adetermined roughly constant wheelconstant speed. The best estimate of the footprint Figure 26 PDFs, shows determined both the PDFs, at a roughly wheel speed. The best estimate of length is the distance between the peaks of the two PDFs curves, that amounts to the distance between the average positions of the trailing and leading edge points.

Sensors 2017, 17, 576

18 of 23

the footprint length is the distance between the peaks of the two PDFs curves, that amounts to the Sensors 2017, 17, 576 18 of 23 Sensors 2017, 17, 576 the average positions of the trailing and leading edge points. 18 of 23 distance between

Figure 26. PDF and polar plot of footprint extrema and plot of local speed over time. Figure 26. PDF and polar plot of footprint extrema and plot of local speed over time Figure 26. PDF and polar plot of footprint extrema and plot of local speed over time

Using the same statistics, Figure 27 shows the PDF curve of the footprint length, whose average Using the same statistics, Figure 27 shows the PDF curve of the footprint length, whose average length, for the analyzed case, case, is is about about 12 12 cm. cm. length, for the analyzed case, is about 12 cm.

Figure 27. Contact length PDF in two different tests. Figure 27. Contact length PDF in two different tests.

The study of the grip phenomena is also possible, and generally very difficult at the actual state The study study of of the the grip grip phenomena phenomena is is also also possible, possible, and generally generally very very difficult at the the actual actual state state difficult of theThe art. Following the analyses presented in [33–34],and it is possible to identify theattransition region of the art. Following the analyses presented in [33–34], it is possible to identify the transition region of the art.grip Following analyses in [33,34],With it is possible to identify the transition region between and slipthe areas using presented special algorithms. the present measurements, we rely only between grip and slip areas using special algorithms. With the present measurements, we rely only between grip and slip areas using special algorithms. With the present measurements, we rely only on the detection based on an abrupt variation on the time history of the first time derivative of the on the detection based on an abrupt variation on the time history of the first time derivative of the strain [33–34]. Figure 28 shows the strain, its first time derivative, and the identified transition abscissa. strain [33–34]. Figure 28 shows the strain, its first time derivative, and the identified transition abscissa.

Sensors 2017, 17, 576

19 of 23

on the detection based on an abrupt variation on the time history of the first time derivative of the Sensors [33,34]. 2017, 17, 576 19 of 23 strain Figure 28 shows the strain, its first time derivative, and the identified transition abscissa.

Figure 28. Summary of the contact patch, transition abscissa and grip/slip area identification. Figure 28. Summary of the contact patch, transition abscissa and grip/slip area identification.

A parameter that is strictly connected to the transition abscissa is the slip ratio. This parameter A parameter that is strictly connected to the transition abscissa is the slip ratio. This parameter is calculated dividing the value of the footprint surface in slip conditions for the value of the whole is calculated dividing the value of the footprint surface in slip conditions for the value of the whole footprint surface. For small slip ratios, i.e., in the case of light torque applied, the transition abscissa footprint surface. For small slip ratios, i.e., in the case of light torque applied, the transition abscissa falls in the neighborhood of the trailing edge, as presented in [33], and the slope change along the falls in the neighborhood of the trailing edge, as presented in [33], and the slope change along the strain curve may become hardly perceptible. In these cases, the experimental identification of the strain curve may become hardly perceptible. In these cases, the experimental identification of the transition abscissa is more problematic, since ambient and numerical noise together with inadequate transition abscissa is more problematic, since ambient and numerical noise together with inadequate time sampling may hide the discontinuity on the first derivative of the strain curve. time sampling may hide the discontinuity on the first derivative of the strain curve. The measured data also allow the estimation of the dissipated power due to hysteresis [35], where The measured data also allow the estimation of the dissipated power due to hysteresis [35], a hysteresis factor is found that explicitly depends on the square of the deformation speed. where a hysteresis factor is found that explicitly depends on the square of the deformation speed. In the end, the data collected in the project will be combined, in forthcoming activities, with some In the end, the data collected in the project will be combined, in forthcoming activities, with advanced models presented in [52,53], to develop a model for the tire structural health monitoring and some advanced models presented in [52–53], to develop a model for the tire structural health damage detection. monitoring and damage detection. 6. Comparative Analysis 6. Comparative Analysis The performances of the developed device are here compared to the most relevant solutions in this The performances of the developed device are here compared to the most relevant solutions in field, which provide state of the art tools for analyzing the contact patch phenomena in a rolling tire. this field, which provide state of the art tools for analyzing the contact patch phenomena in a rolling One of the most acknowledged project is The Cyber™ Tyre [30], developed by Pirelli, which uses tire. a tri-axial accelerometer glued onto the internal carcass of the tire, to estimate the tire-road contact One of the most acknowledged project is The Cyber™ Tyre [30], developed by Pirelli, which dynamics and feed the control logic that governs the active safety of the car. The energy is scavenged uses a tri-axial accelerometer glued onto the internal carcass of the tire, to estimate the tire-road with a nonlinear electro-magnetic device [54]. The energy harvested and the raw acceleration data are contact dynamics and feed the control logic that governs the active safety of the car. The energy is managed by a DSP—Digital Signal Processor. scavenged with a nonlinear electro-magnetic device [54]. The energy harvested and the raw With respect to the Optyre project, the presence of the DSP suggests that (i) the transmission rate acceleration data are managed by a DSP - Digital Signal Processor. and (ii) the sampling frequency of the data are limited by (iii) the power consumption, affecting the With respect to the Optyre project, the presence of the DSP suggests that (i) the transmission rate and (ii) the sampling frequency of the data are limited by (iii) the power consumption, affecting the quality of the measured data. Moreover, (iv) the space resolution of the acquired signal is limited by the intrinsic dimensions of the accelerometer sensor, and finally (v) a radio transmitter should be embedded into the tire. All these limitations are overcome with the proposed technology.

Sensors 2017, 17, 576

20 of 23

quality of the measured data. Moreover, (iv) the space resolution of the acquired signal is limited by the intrinsic dimensions of the accelerometer sensor, and finally (v) a radio transmitter should be embedded into the tire. All these limitations are overcome with the proposed technology. Another related project is based on image processing [32] and has been developed by the Department of Mechanical Engineering of the University of Science in Tokyo. The system relies on the estimation of vertical and frictional loads at the contact interface and uses a CCD camera that monitors the deformation of the internal tire carcass, on which a series of rubber blocks are attached accordingly to a specific pattern. As in the Optyre project, this is an embedded wireless device for the strain measurement, it has the advantage of using a contactless sensor, at the cost of (i) the invasive dimension of the CCD, whose power is supplied by a battery that (ii) needs to be recharged, (iii) substantial modifications of the inner liner caused by the rubber blocks and (iv) a radio receiver needs to be installed. Given the spacing of the rubber blocks and the CCD resolution, (v) the spatial accuracy obtained here is much lower than in Optyre. The same authors also developed a strain monitoring device using a wireless system, through the embedding of a RC oscillating circuit in the inner tire [26]. A portion of the steel wire belt of the tire is the capacitor, and its deformation induces a change in the electrical capacity, which wirelessly permits to identify the tire strain, thanks to a change in the frequency response of the oscillating circuit. With respect to the Optyre technology a clear advantage of this device is to use the carcass of tire as a sensor, nevertheless with this solution (i) the measurement is averaged on a large tire surface, that is not negligible. Moreover, (ii) the frequency bandwidth of the oscillating circuit limits the maximum operating speed, finally, (iii) road tests are not reported. The CAIS-Contact Area Identification System [31], developed by Bridgestone, is able to estimate the contact patch, the load and side force, thanks to direct measurements of strain, acceleration, pressure and temperature, with a wireless sensor placed inside the tire. The system can also classify the road profile and estimate the tire tread wear. This solution may allow real-time testing but, as mentioned in [31], (i) an energy harvester is needed and not designed yet. The data processing algorithms, on which this technology is based, rely on the assumption that the highest speed of deformation occurs when the monitored points reach the edges of the contact patch. About this hypothesis, the authors showed in [33] that using the formulas presented in [31], the contact patch is underestimated. It was indeed demonstrated, theoretically and experimentally, that (ii) the extrema of the deformation speed fall inside the contact patch [33]. In conclusion, although Optyre presents some important advantages with respect to aforementioned technologies, such as the direct strain measurement with negligible insertion errors, a high spatial resolution, a multiple points measurement within a single fiber, high strain sensitivity, high frequency measurements and the absence of any sort of power source within tire, it is still a prototypal device that needs to be engineered for a widespread commercial use. 7. Conclusions The experimental apparatus developed in the Optyre project has been presented in this paper. The innovative experimental system relies on the embedding of optical lines with FBG sensors within the tire carcass, together with the use of other sensors, such as a phonic wheel, a uniaxial accelerometer, a dynamic temperature sensor and a GPS, which allow the real-time identification of key parameters, such as the tire stress during rolling, the residual grip and the rolling resistance, as shown in previous papers and will be presented in forthcoming publications, which will be focused on both the development and the engineering of a vehicle dynamics control system based on the presented hardware. The validity of the developed smart tire was investigated with an extensive experimental campaign, and by comparison with other existing technologies. Results confirm the potential of the present apparatus, which provides signals with high resolution and accuracy, does not require

Sensors 2017, 17, 576

21 of 23

energy harvesting or complex transmission systems, and is capable of long-term acquisitions under adverse conditions. The reliability of the experimental setup developed in the Optyre project suggests moving towards new research directions, in order to investigate further phenomena involving the tire revolution: the study of the impact dynamics of the leading edge applying the model presented in [55]; the evaluation of the tire-road dynamics in presence of interstitial fluids, such as in aquaplaning conditions, following the thesis developed in [56]; the development of a new high order constitutive relationship between stress and strain, for the complex structures of the tire, through models close to the ones suggested in [57]; or the use of the theory presented in [58] to evaluate a richer phenomenology of the contact between rough surfaces. Acknowledgments: The research work was supported by the Department of Mechanical and Aerospace Engineering, Sapienza, University of Rome, Via Eudossiana 18, 00184 Rome, Italy, and CNIT—Consorzio Nazionale Interuniversitario per le Telecomunicazioni, Via Gian Paolo Usberti 181/A, 43124 Parma, Italy, in the context of the Research Project ITS 2020. Author Contributions: Gianluca Pepe and Nicola Roveri developed the design of the setup together with the prototype realization. Francesco Coppo developed the data processing to extract the physical information from the acquired signals, and Antonio Carcaterra is the scientific supervisor and the director of the Vehicle Dynamics and Mechatronics Lab. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4. 5.

6.

7. 8. 9. 10. 11.

12. 13. 14.

Intelligent Tyre Systems—State of the Art and Potential Technologies; Technical Research Centre of Finland (VTT): Espoo, Finland, 2003. Jo, H.Y.; Yeom, M.; Lee, J.; Park, K.; Oh, J. Development of Intelligent Tire System; No. 2013-01-0744; SAE Technical Paper: Warrendale, PA, USA, 2013. Al-Mola, M.H.; Mailah, M.; Samin, P.M.; Muhaimin, A.H. Hybrid Control Scheme for Pursuing Performance of an Anti-Lock Brake System. Int. Rev. Model. Simul. 2013, 6, 1961–1967. Kubba, A.E.; Jiang, K. A comprehensive study on technologies of tyre monitoring systems and possible energy solutions. Sensors 2014, 14, 10306–10345. [CrossRef] [PubMed] Singh, K.B.; Arat, M.A.; Taheri, S. Development of a smart tire system and its use in improving the performance of a collision mitigation braking system. In Proceedings of the ASME 2012 International Mechanical Engineering Congress and Exposition, Houston, TX, USA, 9–15 November 2012; pp. 77–87. Erdogan, G.; Borrelli, F.; Tebano, R.; Audisio, G.; Lori, G.; Sannazzaro, J. Development of a New Lateral Stability Control System Enhanced with Accelerometer Based Tire Sensors. In Proceedings of the ASME 2010 Dynamic Systems and Control Conference, Cambridge, MA, USA, 13–15 September 2010; pp. 841–848. Cheli, F.; Sabbioni, E.; Sbrosi, M.; Brusarosco, M.; Melzi, S. Enhancement of ABS Performance through On-Board Estimation of the Tires’ Response by Means of Smart Tires; SAE Technical Paper: Warrendale, PA, USA, 2011. Niskanen, A.; Tuononen, A.J. Three Three-Axis IEPE Accelerometers on the Inner Liner of a Tire for Finding the Tire-Road Friction Potential Indicators. Sensors 2015, 15, 19251–19263. [CrossRef] [PubMed] Yilmazoglu, O.; Brandt, M.; Sigmund, J.; Genc, E.; Hartnagel, H.L. Integrated InAs/GaSb 3-D magnetic field sensors for ‘the intelligent tire’. Sens. Actuators A Phys. 2001, 94, 59–63. [CrossRef] Yi, J. A piezo-sensor based ‘smart tire’ system for mobile robots and vehicles. IEEE/ASME Trans. Mechatron. 2008, 13, 95–103. [CrossRef] Yu, H.; Taheri, S.; Duan, J.; Qi, Z. An Integrated Cooperative Antilock Braking Control of Regenerative and Mechanical System for a Hybrid Electric Vehicle Based on Intelligent Tire. Asian J. Control 2016, 18, 55–68. [CrossRef] Intelligent Tyre for Accident-free Traffic. Available online: http://virtual.vtt.fi/virtual/proj3/apollo/index. htm (accessed on 10 January 2017). Matsuzaki, R.; Kamai, K.; Seki, R. Intelligent tires for identifying coefficient of friction of tire/road contact surfaces using three-axis accelerometer. Smart Mater. Struct. 2014, 24, 025010. [CrossRef] Hannan, M.A.; Hussain, A.; Samad, S.A. System interface for an integrated intelligent safety system (ISS) for vehicle applications. Sensors 2010, 10, 1141–1153. [CrossRef] [PubMed]

Sensors 2017, 17, 576

15. 16. 17. 18. 19. 20. 21. 22.

23. 24. 25.

26. 27. 28. 29.

30. 31. 32.

33. 34. 35.

36. 37. 38.

22 of 23

Arat, M.; Taheri, S. Identification of Road Surface Friction for Vehicle Safety Systems; SAE Technical Paper: Warrendale, PA, USA, 2014. Edelmann, J.; Gobbi, M.; Mastinu, G.; Ploechl, M.; Previati, W.G. Friction estimation at tire-ground contact. SAE Int. J. Commer. Veh. 2015, 8, 182–188. [CrossRef] Zhang, Y.; Yi, J.; Liu, T. An embedded local force sensor for insitu tire-road interaction measurements. IEEE Sens. J. 2013, 13, 1756–1765. [CrossRef] Erdogan, G.; Alexander, L.; Rajamani, R. Estimation of tire-road friction coefficient using a novel wireless piezoelectric tire sensor. IEEE Sens. J. 2011, 11, 267–279. [CrossRef] Pohl, A.; Steindl, R.; Reindl, L. The “Intelligent tire” utilizing passive saw sensors measurement of tire friction. IEEE Trans. Instrum. Meas. 1999, 48, 1041–1046. [CrossRef] Zhang, Y.; Yi, J. Static Tire/Road Stick–Slip Interactions: Analysis and Experiments. IEEE/ASME Trans. Mechatron. 2014, 19, 1940–1950. [CrossRef] Matsuzaki, R.; Todoroki, A. Passive wireless strain monitoring of actual tire using capacitance-resistance change and multiple spectral features. Sens. Actuators A Phys. 2006, 126, 277–286. [CrossRef] Savaresi, S.M.; Tanelli, M.; Langthaler, P.; Del Re, L. New regressors for the direct identification of tire deformation in road vehicles via ‘intire’ accelerometers. IEEE Trans. Control Syst. Technol. 2008, 16, 769–780. [CrossRef] Guan, H.; Duan, C.; Lu, P.; Zhan, J.; He, Y. The Research of Tire Mechanics at Lower-Speed for Interactive Developing; SAE Technical Paper: Warrendale, PA, USA, 2015. Gobbi, M.; Previati, G.; Mastinu, G. Refined design of a measuring wheel. In Proceedings of the ASME Design Engineering Technical Conference IDETC/CIE, Washington, DC, USA, 28–31 August 2011. Matsuzaki, R.; Hiraoka, N.; Todoroki, A.; Mizutani, Y. Optical 3D Deformation Measurement Utilizing Non-planar Surface for the Development of an “Intelligent Tire”. J. Solid Mech. Mater. Eng. 2010, 4, 520–532. [CrossRef] Matsuzaki, R.; Todoroki, A. Wireless monitoring of automobile tires for intelligent tires. Sensors 2008, 8, 8123–8138. [CrossRef] [PubMed] Kim, S.J.; Kim, K.S.; Yoon, Y.S. Development of a tire model based on an analysis of tire strain obtained by an intelligent tire system. Int. J. Automot. Technol. 2015, 16, 865–875. [CrossRef] Yang, X.; Olatunbosun, O.; Garcia-Pozuelo, D.; Bolarinwa, E. FE-Based Tire Loading Estimation for Developing Strain-Based Intelligent Tire System; SAE Technical Paper: Warrendale, PA, USA, 2015. Yang, X.; Olatunbosun, O.; Ramos, D.G.P.; Bolarinwa, E. Experimental Investigation of Tire Dynamic Strain Characteristics for Developing Strain-Based Intelligent Tire System. SAE Int. J. Passeng. Cars-Mech. Syst. 2013, 6, 97–108. [CrossRef] Audisio, G.; Cheli, F.; Melzi, S.; Velardocchia, M. Cyber™ Tyre for Vehicle Active Safety. In Proceedings of the AIMETA 2009 Congress, Ancona, Italy, 14–17 September 2009. Morinaga, H.; Wakao, Y.; Hanatsuka, Y.; Kobayashi, A. The possibility of intelligent tire. In Proceedings of 31st FISITA Automotive, Congress, Yokohama, Japan, 22–27 October 2006; p. F2006V104. Matsuzaki, R.; Hiraoka, N.; Todoroki, A.; Mizutani, Y. Strain Monitoring and Applied Load Estimation for the Development of Intelligent Tires Using a Single Wireless CCD Camera. J. Solid Mech. Mater. Eng. 2012, 6, 935–949. [CrossRef] Roveri, N.; Pepe, G.; Carcaterra, A. OPTYRE—A new technology for tire monitoring: Evidence of contact patch phenomena. Mech. Syst. Signal Process. 2016, 66, 793–810. [CrossRef] Carcaterra, A.; Roveri, N. Tire grip identification based on strain information: Theory and simulations. Mech. Syst. Signal Process. 2013, 41, 564–580. [CrossRef] Roveri, N.; Pepe, G.; Carcaterra, A. On line estimation of rolling resistance for intelligent tires. In Proceedings of the International Conference on Noise and Vibration Engineering (ISMA 2016), Leuven, Belgium, 19–21 September 2016. Carcaterra, A.; Roveri, N.; Platini, M. Device and Method for Optically Measuring the Grip of a Tyre and Tyre Suitable for Said Measuring Property; University of Rome “La Sapienza”: Roma, Italy, 2011. Gladwell, G.M.L. Inverse Problems in Vibration; Springer: Berlin/Heidelberg, Germany, 2004. Antunes, P.; Dominguez, F.; Granada, M.; André, P. Mechanical Properties of Optical Fibers, Selected Topics on Optical Fiber Technology; Yasin, M., Ed.; InTech: Vienna, Austria, 2012.

Sensors 2017, 17, 576

39. 40. 41. 42. 43. 44. 45. 46.

47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57.

58.

23 of 23

SmartScan Interrogator Datasheet. Available online: http://www.smartfibres.com/docs/SmartScan.pdf (accessed on 17 November 2016). Werneck, M.M.; Allil, R.C.S.B.; Ribeiro, B.A.; de Nazaré, F.V.B. A Guide to Fiber Bragg Grating Sensors: Current Trends in Short- and Long-period Fiber Gratings; INTECH: Vienna, Austria, 2013. High Temperature Fiber Bragg Grating Datasheet. Available online: http://www.technicasa.com/www/ wwwimg/Uploads/files/Technica%20High%20Temp%20FBG-160106.pdf (accessed on 17 November 2016). Corning® SMF-28e+® Optical Fiber Product Information. Available online: http://www.corning.com/ media/worldwide/coc/documents/PI1463_07-14_English.pdf (accessed on 17 November 2016). Single Channel Fiber Rotary Joint. Available online: http://www.princetel.com/forj_rpt.asp (accessed on 10 January 2017). LMS Scadas Catalogue. Available online: http://www.dta.com.tr/pdf/lms_scadas/LMS_Scadas_veri_ toplama_sistemleri.pdf (accessed on 10 January 2017). Elen AA020020 Sensor Datasheet. Available online: http://www.elensrl.com/it/fileuploader/download/ download/?d=0&file=custom%2Fupload%2FFile-1412090743.pdf (accessed on 17 November 2016). Optris Thermometer Datasheet. Available online: http://www.optris.com/optris-csmicro-lt?file=tl_ files/pdf/Downloads/Compact%20Series/Data%20Sheet%20optris%20CSmicro%20LT.pdf (accessed on 10 January 2017). Telemetry System Datasheet. Available online: https://drive.google.com/a/uniroma1.it/file/d/ 0B5RAK5cV5xPKMk1mTzRmMVN3UzA/view?usp=sharing (accessed on 10 January 2017). Hoffmann, K. An Introduction to Measurements Using Strain Gages; Hottinger Baldwin Messtechnik GmbH: Darmstadt, Germany, 1989. Carcaterra, A.; Roveri, N.; Pepe, G. Fractional dissipation generated by hidden wave-fields. Math. Mech. Solids 2015, 20, 1251–1262. [CrossRef] Roveri, N.; Carcaterra, A.; Akay, A. Frequency intermittency and energy pumping by linear attachments. J. Sound Vib. 2014, 333, 4281–4294. [CrossRef] Genta, G. Meccanica Dell’autoveicolo; Levrotto e Bella: Torino, Italy, 1982. Roveri, N.; Carcaterra, A.; Sestieri, A. Real-time monitoring of railway infrastructures using fibre Bragg grating sensors. Mech. Syst. Signal Process. 2015, 60, 14–28. [CrossRef] Roveri, N.; Carcaterra, A. Unsupervised identification of damage and load characteristics in time-varying systems. Contin. Mech. Thermodyn. 2015, 27, 531–550. [CrossRef] Magneto-Mechanical Energy Harvester for Wireless Sensors in Tyre, Shoes and Other Components. Available online: http://www-2.unipv.it/compmech/seminars/bonisoli-pres.pdf (accessed on 10 January 2017). Andreaus, U.; Placidi, L.; Rega, G. Soft impact dynamics of a cantilever beam: Equivalent SDOF model versus infinite-dimensional system. Proc. Inst. Mech. Eng. Part C 2011, 225, 2444–2456. [CrossRef] Dell’isola, F.; Madeo, A.; Seppecher, P. Boundary conditions at fluid-permeable interfaces in porous media: A variational approach. Int. J. Solids Struct. 2009, 46, 3150–3164. [CrossRef] Dell’Isola, F.; Seppecher, P. The relationship between edge contact forces, double forces and interstitial working allowed by the principle of virtual power. Comptes Rendus l’Academie Sciences Series IIb Mecanique Physique Chimie, Astronomie 1995, 321, 303–308. Misra, A.; Huang, S. Micromechanics based stress-displacement relationships of rough contacts: Numerical implementation under combined normal and shear loading. Comput. Model. Eng. Sci. 2009, 52, 197–215. © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).