A Sensor for the Optical Detection of Dangerous Road Condition

Armando Piccardi and Lorenzo Colace

NooEL - Nonlinear Optics and OptoElectronics Laboratory, Department of Engineering, University of Study "ROMA TRE",

Via Vito Volterra 62, Rome, Italy

Keywords: Optical Sensor, Light Scattering, Polarization Contrast Ratio.

Abstract: We present the design, realization and characterization of a sensor for the optical detection of hazardous

road condition. The device exploits the radiation from a incoherent source to detect the polarized

components of the light scattering from a rough surface and calculate a signal dependent on the surface

state. We propose two distinct geometries, working with forward and backward scattering components,

assessing the device performance in terms of reliability and compactness. In both cases, the sensor allows to

discriminate potentially dangerous states like the presence of water (either wet surface or covered by a

layer) or ice on an asphalt sample, in opposition to the dry surface representing a safe condition.

1 INTRODUCTION

The automotive sector is currently characterized by a

high demand of sensing devices either for the

vehicle performances evaluation or the assessment

of the drivers safety (Guerrero-Ibáñez, 2018). With

reference to the latter, a number of methods and

devices have been proposed in the last years in order

to detect the presence of atmospheric agents on the

road surface. In particular, ice detection rises great

attention because of it is not limited to the

automotive sector, but it cover a range of possible

application ranging from aircraft security to

monitoring of surfaces under extreme conditions

(Flatscher, 2017; Amiropoulos, 2018; Muñoz, 2016).

Nevertheless, the trade off between reliability

and costs if often a limiting factor for a device

thought to meet the market requirements: contact

sensors for example, have high maintenance costs

though they result quite reliable (Zhi, 2015; Roy,

2000). On the contrary, contact-less devices

(Ikiades, 2007) often suffer of poor selectivity to

different agents.

The most used methods exploit capacitive,

piezoelectric or electrical elements to detect

temperature and humidity and evaluate the presence

of ice (Wollenweber, 2018; Tabatabai, 2017).

Optical methods and sensors encompass fiber-

optics, devices based on infrared image processing

light reflection and diffusion at different wavelength

ranges (Abdel-Moati, 2018; Finkele, 1997; Ogura,

2002). In particular, light scattering polarization

have been recently used to detect the presence of ice

or water on asphalt surfaces (Colace, 2013).

In this work, we present the implementation of a

device employing an optical method based on the

detection of the polarization components of infrared

light scattering on asphalt surfaces. Evaluating the

ratio between two orthogonally polarized

components, we manage to distinguish between

several surface conditions and thus between a safe

and potentially dangerous states.

The manuscript is organized as follows: first, the

optical method, the devices components and the

parameters needed to characterize the system are

presented. Then, we will describe the realization of

different versions of the sensor, changing the

geometric arrangement and number of components.

The last section is dedicated to the results and the

discussion about the system performances.

2 PRINCIPLE OF OPERATION

The proposed sensor works with a light source

incident on a sample (asphalt) surface, where the

diffusion process changes the distribution of

orthogonally (TE and TM) polarized components of

the scattered light. In this way, we can obtain

information about the surface under investigation

depending on the detection of the two polarization

components.

Piccardi, A. and Colace, L.

A Sensor for the Optical Detection of Dangerous Road Condition.

DOI: 10.5220/0007572002830288

In Proceedings of the 7th International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS 2019), pages 283-288

ISBN: 978-989-758-364-3

283

In order to implement this method, we need a light

source, some detectors and the electronics to process

the acquired signals and extract the suitable figures

of merit. As done in (Colace, 2013), a useful

parameter is the Polarization Contrast Ratio (PCR),

defined by the difference between the intensities of

the two polarization components (I

, I

)

normalized by the sum of the two:

TMTE







(1)

Figure 1: Block diagram of the sensor design. The

modulate source (S) emits radiation collimated by a lens

(L). The scattered light is detected by the three

photodiodes: two (PD1 and PD2) collecting orthogonally

polarized light by employing two polarizers (P1 and P2),

and the third (PD3) detecting the unpolarized light. Trans

impedance amplifiers (TIA) convert the signals to be

processed by the lock-in amplifier. The resulting voltage

values allow the calculation of the PCR.

Moreover, in our work we also considered the

total unpolarized scattering as a parameter to be

measured.

The employed source is a LED operating at

λ=980nm. The chosen wavelength stands in a

spectrum range where polarized scattering is rather

sensitive to material properties; moreover, near

infrared radiation prevents direct disturbance of

human eyes while driving and it is compatible with

reliable and widely available detection components,

in the framework of low cost device realization. A

lens is used to compensate for the source diffraction

and to regulate the spot size on the sample in order

to optimize the detection.

In order to detect the polarized components and

extract the PCR we need two photodetectors, each

one sensitive to one component by employing two

orthogonal polarizers. The detectors are provided

with low pass filters screening any optical

disturbances from ambient light. A third

photodetector is employed to collect the unpolarized

scattering and provide additional formation about

the surface.

The reading electronics converts the current

signals into voltage signals through distinct trans-

impedance amplifiers (TIA) and performs a

homodyne detection through a lock-in amplifier. For

this reason, the source driver is designed in order to

obtain a modulated signal at f=1kHz. The voltage

value carrying the information about the polarization

components of the scattered light are finally ready to

be processed in order to obtain the PCR. We stress

that the latter is calculated as in eq.1, but using the

voltage values corresponding to the intensities of the

polarized components.

The complete set up is sketched in fig. 1.

3 GEOMETRIC

ARRANGEMENTS

Given the working principle of the sensor, the

components and the quantities needed for the surface

detection, we can arrange the set-up in different

ways. The sensor works with scattered light, which

is distributed over a cone depending on the kind of

surface and its roughness as well. For a smooth

sample, we expect the incident radiation to be almost

reflected forward, while for a rough surface, the

light scattering can be supposed isotropic. Thus, the

choice of the position where to detect the scattered

light can affect the system performances. We

investigated two main geometric configurations: the

first has the source and the detectors in a θ-2θ

arrangement, thus detecting the mirror-like polarized

reflections; the second works with backscattering,

thus detecting the radiation diffused backward and it

has the source and the photodiodes arranged in the

same position.

The former geometry allow to operate with

higher signal levels, but it has higher costs linked to

the large area to be dedicated to the sensor. The

latter required higher sensitivity as the backscattered

signals can be considerably lower but is more

convenient in terms of compactness.

PHOTOPTICS 2019 - 7th International Conference on Photonics, Optics and Laser Technology

284

Figure 2: Sketch of the two considered geometries. (a) The

mirror-like configuration exploits the Snell law to detect

the forward scattered light. (b) The backscattering

configuration works with backward diffusion. For both the

configurations, a third photodiode can be added to detect

the unpolarized scattering.

Moreover, the presence of a third photodetector,

collecting the unpolarized scattering to be used as a

further measurement to discriminate between

different surfaces is evaluated. Its position could be

varied as well, but as a reference position we will

consider the direction perpendicular to the surface

normal.

The geometrical arrangements are resumed in

the sketch of fig.2.

4 RESULTS AND DISCUSSION

For both the geometries, the characterization of the

sensor is divided into two parts: a preliminary

measurements is done by varying the incidence

angle of the beam, to check the dependence of the

PCR on the source orientation (Videen, 1992). The

polarization of the source is varied as well, in order

to further improve the system sensitivity to different

conditions, exploiting the polarization dependence

of the Fresnel coefficients. Then, once optimized the

geometry, the detection of the polarization

components and the calculation of the suitable

parameters are performed on asphalt samples.

In all the presented measurements the source has

a power of 18mW and it is at a distance of about

30cm from the surface. We stress that the employed

values can obviously be scaled due to the peculiar

application, but at the moment they try to reproduce

a possible situation when mounting the device on a

moving vehicle or on a road side. Several conditions

have been considered. The safe condition is

indicated by the values provided by dry asphalt,

corresponding to the reference state. Then other

surfaces have been prepared corresponding to

different atmospheric or road conditions, linked to

different level of hazard: wet asphalt, surface

covered by a water layer and the presence of an ice

layer.

4.1 Mirror-like Configuration

For this configuration the source and the detectors

are mounted on a goniometric system ensuring that

incidence and detection angle are equal. Incidence

angle have been varied between 20° and 60° with

respect to the surface normal. Fig. 3 shows the PCR

versus incidence angle for some of the surface

conditions, when the source radiation is unpolarized

or horizontally/vertically polarized.

As it can be seen, both the angle dependence

and the absolute value of PCR are strongly

dependent on polarization. For unpolarized or

horizontally polarized radiation the curves results

superposed, for almost all the angular range. On the

other hand, vertical polarized light allows the

discrimination of different conditions, in particular,

between 50° and 60°. Moreover, for 50° the dry

asphalt (safe) case has the highest value of PCR,

thus it is well distinguishable from the other

conditions. For these reasons, we choose θ=50° to

complete the characterization.

At this angle, we calculate the PCR and plot it

versus the unpolarized scattering. We performed the

measurements a number of times to include the

statistical errors of both measurements, also

considering several samples for each conditions.

Thus, we obtained dispersion graphs on the

parameter space for all the road condition, as

reported in fig.4. In this way, the separation of the

clouds allows finding portions of the parameter

space associated to one state, while clouds

superpositions indicate the possibility of obtaining a

measurement with a certain error percentage.

Fig.4(a) shows a good separation between the

states but for the couple dry-ice, for which we

calculate an error of about 10%. This represents a

problem because the system is interpreting a

potentially dangerous condition as a safe one (or

vice versa). For this reason we checked if it exists

another way to represent the different road

A Sensor for the Optical Detection of Dangerous Road Condition

285

Figure 3: Mirror-like configuration. PCR as a function of

the incidence angle for (a) unpolarized, (b) horizontally

polarized and (c) vertical polarized incident beam.

conditions: we substitute the unpolarized scattering

measurement with the sum of the two polarized

components, so that we plot PCR versus sum.

This choice has the further advantage of

avoiding the presence of the third photodiode,

reducing the complexity and the size of the whole

sensor. Results are shown in fig.4(b).

In the new data representation the different states

are all distinguishable, allowing a good evaluation of

the asphalt conditions. Even if the wet and water

cases results close to each other, in practice they

correspond to similar hazards and thus their wrong

evaluation would have no consequences on the

system performances.

Figure 4: Mirror-like configuration. (a) PCR versus

Unpolarized scattering (V

). (b) PCR versus the sum of the

polarization component.

4.2 Backscattering Configuration

The second investigated geometry encompasses

source and detectors on the same side, i.e. it detects

the backscattering from the asphalt surfaces. We

carry out the characterization as for the previous

geometry: first, we identify the optimum sensor

orientation, by measuring the PCR as a function of

the incidence angle for three different polarizations

of the source (unpolarized, horizontal and vertical),

as reported in fig.5.

This geometry allows a better distinction among

the road states than the previous one. Nevertheless,

once again the vertical polarization provides the best

results; in particular, at θ=20° the curves are well

distributed over a quite range of PCR, thus the

complete characterization is performed at this angle.

We plotted the dispersion graphs in both the case

with or without the third photodiodes, i.e.

considering either the unpolarized scattering or the

sum of the orthogonally polarized polarization. In

this case, as visible for the direct comparison

between panels (a) and (b) of fig.6, differences in the

position and distributions of the clouds

corresponding to the road states are negligible: this

means that the two parameters are equivalent and the

PHOTOPTICS 2019 - 7th International Conference on Photonics, Optics and Laser Technology

286

Figure 5: Backscattering radiation. PCR as a function of

the incidence angle for (a) unpolarized, (b) horizontally

polarized and (c) vertical polarized incident beam.

third photodiode is not necessary at all, thus can be

removed with improved sensor compactness and

complexity. Moreover, no superposition of clouds is

present. Thus, it is straightforward to identify the

threshold values and functions for the evaluation of

the surface state, as we do in fig.6(b) where we

proposed an example of space division based on

lines.

With this kind of division the sensor is fully

characterized: when the system is working, the

measurement of PCR and sum of polarized

scattering will generate a point falling into one of the

portions of the parameter space. Thus, a

programmed logic can be implemented to process

the data and give the corresponding indication of the

current road state.

Figure 6: Backscattering configuration. (a) PCR versus

Unpolarized scattering (V

). (b) PCR versus the sum of the

polarization component.

5 CONCLUSIONS

We designed, realized and characterized a sensor

able to distinguish among for different conditions -

one safe and three corresponding to different

potentially dangerous situations - of an asphalt

surface, in the framework of road security. The

optical method used for the detection of a surface

state is based on the measurement of the polarization

components of the scattered radiation. The

investigation of different geometries and

configurations allowed to find the optimal set-up in

terms of costs and compactness, and to discriminate

the presence of water or ice on asphalt surfaces with

good reliability. Though in a preliminary version,

the realized device and the detection principle are

promising and deserve further investigation.

Regarding the future developments, the work on the

sensor includes the realization of the logic to process

the raw data and automatically give information

about the surface state and the investigation of other

kind of conditions; on the other hand, statistics about

the raw data are the object of numerical

investigation, in order to improve the detection

reliability, as well as a study on the different textures

of the investigated surfaces.

A Sensor for the Optical Detection of Dangerous Road Condition

287

ACKNOWLEDGEMENTS

This work was supported by ACTPHAST (Access

CenTer for PHotonics InnovAtion Solutions and

Technology Support), grant number P2016-29

(SODARC - Sensor for the Optical detection of

DAngerous Road Conditions).

REFERENCES

Abdel-Moati, H., Morris, J., Zeng, Y., Wesley Corie, M.,

Garas Yanni, V., 2018. Near field ice detection using

infrared based optical imaging technology. Optics &

Laser Technology, 99, 402-410.

Amiropoulos, K., Spasopoulos, D., Ikiades, A., 2018.

Fiber optic sensor for ice detection on aerodynamic

surfaces using plastic optic fiber tapers. Advanced

Photonics (Sensors), OSA Technical Digest, SeM4E.6.

Colace, L., Santoni, F., Assanto, G., 2013. A near-infrared

optoelectronic approach to detection of road

conditions. Optical and Laser Engineering, 51, 633-

636.

Finkele, R., 1997. Optical road-ice detector operating in

the near infrared. Electronic Letters, 33, 1153-1154.

Flatscher, M., Neumayer, M., Bretterklieber, T., 2017.

Maintaining critical infrastructure under cold climate

conditions: a versatile sensing and heating concept.

Sensors and Actuators A: Physical, 267, 538-546.

Guerrero-Ibáñez, J., Zeadally, S., Contreras-Castillo, J.,

2018. Sensor Technologies for Intelligent

Transportation Systems. Sensors, 18, E1212.

Ikiades, A., Howard, G., Armstrong D.J. et al., 2007.

Measurement of optical diffusion properties of ice for

direct detection ice accretion sensors. Sensors and

Actuator A. 140, 24-31.

Muñoz, C., Márquez, F., Tomás, J., 2016. Ice detection

using thermal infrared radiometry on wind turbine

blades. Measurement, 93, 157-163.

Ogura, T., Kageyama, I., Katsuhisa N., et al., 2002. Study

on a road surface sensing system for snow and ice

road. Journal of Society of Automotive Engineers

Review, 23, 333-339.

Roy, S., De Anna, R.G., Mehregani, M., Zakar, E., 2000.

A capacitive ice detection microsensor. Sensor

Materials, 12, 1-14.

Tabatabai, H., Aljuboori, M., 2017. A Novel Concrete-

Based Sensor for Detection of Ice and Water on Roads

and Bridges. Sensors, 17, E2912.

Videen, G., Hsu, J., Bickel, W.S., Wolfe, W.L., 1992.

Polarized light scattered from rough surfaces. Journal

of Optical Society of America, 7, 1111-1118.

Wollenweber, G. C., 2018. US Patent App., 15/275, 013.

Zhi, X., Cho, H.C., Wang, B., Ahn, C.H., Moon, H.S., Go,

J.S., 2015. Development of a Capacitive Ice Sensor to

Measure Ice Growth in Real Time. Sensors, 5, 6688-

6698.

PHOTOPTICS 2019 - 7th International Conference on Photonics, Optics and Laser Technology

288