|
1.INTRODUCTIONThe Autonomous driving technology of vehicles plays an important role in military and civil affairs, and which can undertake tasks such as explosive removal and de-mining, material transportation, and automobile driving assistance. It has developed rapidly in recent years, and represents an important development direction of automobile technology1. The core of autonomous vehicle driving technology is driving safety2. Most of the existing autonomous driving schemes adopt microwave radar or lidar to detect obstacles, and use cameras to identify pedestrian and road traffic signal signs3-4. There are few studies on the use of laser sensors to perceive road information. Different road surface has different physical properties and covering characteristics, such as asphalt, cement, soil road surface and icing, snow, water and so on. The above factors have an important influence on the choice of vehicle speed and braking distance. The laser plays an important role in space target detection and location. At the laser wavelength scale, the road surface can be regarded as random rough surface. Which can modulate the incident polarized light, and the scattered light contains roughness information. The research on the characteristics of laser scattering on road surface and the perception of road surface information can provide decision support for autonomous driving of vehicles5 have important application value. Many scholars have carried out relevant studies on the modeling of the electromagnetic characteristics of road surface. Liu6 carried out studies on the dielectric constant and scattering coefficient of asphalt and cement road surface. Zhang7 studied the surface-volume composite electromagnetic scattering model of millimeter band cement and asphalt road surface, and used the measured surface scattering data of multi-frequency band to jointly invert the surface equivalent parameters. Zhu8 used Finite Difference Time Domain (FDTD) method to study the cement concrete road surface system with one-dimensional band-limited Weierstrass fractal characteristics. Cao9 carried out a study on the composite electromagnetic scattering characteristics of target and road surface environment. Foreign scholar Kamal Sarabandi10-13 conducted a series of in-depth studies on millimeter-wave scattering on asphalt and cement road surface, et al. Most of the above researches focus on electromagnetic scattering characteristics of microwave and millimeter-wave road surface, while there are few reports on laser scattering characteristics of cement road surface with millimeter fluctuation. In this paper, we adopted virtual instrument technology to design the control system. The Labview software communicated with MCU through serial port bus. The MUC controlled the stepper motor drivers to rotate the two dimensional turntable. The MCU obtained the scattered light intensity from the photoelectric sensor via I2C bus. The full angular measurement of hemispheric space scattering light is realized. 2.LASER SCATTERING MEASUREMENT SYSTEM2.1Optical path designThe Stokes vector S = [S0, S1, S2, S3]T contains four components. S0 represents the total intensity of scattered light. S1 represents the difference between the vertical and parallel components of the scattered light. S2 represents the difference between ±45°. S3 represents the difference between left and right circularly polarized component. The optical path design is shown in Figure 1. The semiconductor laser with wavelength 650nm is chosen as the light source, and it has good coherence and high stability. It is a continuous laser containing an expanding beam and a collimating lens. The polarizing film can further improve the polarization degree of the emitting laser. Daheng Photoelectric GCL-05 artificial polarizing film is adopted. It’s pass-through aperture is 25.4mm. The extinction ratio is 500:1, and the applicable band range is 400-700nm. The polarization direction of GCL-05 polarizer is not exactly marked, so it needs to be calibrated again. A quarter wave plate is used to adjust the polarization of the incident light. By adjusting the fast axis, the linearly polarized light passing through a quarter wave plate can become circularly polarized light. By rotating the fast axis, the half wave plate can adjust the vibration direction of the incident light. Four groups of incident completely polarized light can be modulated by the above devices. They are I(0°,0), I(90°,0), I(45°,0), I(45°, π / 2). The I(0°,0) is the intensity of light scattered parallel to the plane of incidence. The I(90°,0) is the intensity of light scattered perpendicular to the plane of incidence. The I(45°,0) is the intensity of the scattered light at an angle of 45º from the plane of incidence. The I(45°, π / 2) is right-handed circularly polarized light. The measurement formula of Stokes vector is 2.2Mechanical structure designIn order to study the laser scattering characteristics of cement road surface with different roughness, a high-precision laser scattering measurement system is designed, as shown in Figure 2. The system is based on virtual instrument technology, and the Labview software is used to design flexible and expandable instrument panel relying on the computer hardware platform. Labview realizes serial port communication with micro controller unit (MCU) through graphical programming, and issues control commands to MCU. The single chip microcomputer controls the stepper motor 1 to realize accurate adjustment of the angle of the turntable, and controls the stepper motor 2 to rotate the screw drive mechanism to drive the sample to be tested to move. The average value of the scattered light intensity can be obtained by multiple measurements. 2.3Control systemA high-precision digital light intensity detection module with BH1750FVI as the core was made. The BH1750FVI is a 16-bit digital output integrated circuit. It uses IIC bus to communicate with single chip microcomputer to upload measurement data. Its internal photoelectric sensor has optical efficiency similar to that of human eyes, with a peak sensitive wavelength of 560nm, which can achieve a wide range of optical intensity measurement within 65535lx. In order to avoid the situation that scattered light intensity is weak and difficult to measure, LD-650Y2 high-power semiconductor laser is adopted for laser emission, with output power up to 250mW and working voltage of 3-4.5V. Output power can be adjusted with working voltage. The transmitter module contains focusing and collimating lens to realize adjustable laser speckle diameter. The schematic diagram of control system as shown in Figure 3. The Labview running on the upper computer issues the control command of stepper motor 1 to the MCU through serial port communication, so that the turntable rotates to the specified angle within the range of 0°-180°. The Labview issues the stepper motor 2 control command to the MCU to make the screw drive mechanism move the sample to be tested. The MCU issues the data acquisition command. The MCU communicates with the BH1750FVI module through IIC bus, and obtains the measurement data. It uploates the measured data to the upper computer LABVIEW. Cycle for the next data acquisition. 3.THE EXPERIMENTAL RESULTS AND DISCUSSIONThe incident angle is 30°. The wavelength, correlation length and height of incident light represent the optical roughness of rough surface. The larger the incident wavelength and height root mean square (RMS), the smaller the correlation length, and the rougher the roughness surface. We made four cement blocks, and the measured the root-mean-square values are as follows: A- δ=0.211mm, B- δ=0.302mm, C- δ=0.407mm, D- δ=0.512mm. 3.1The Stokes vector of the scattered lightThe experiment results of scattered light Stokes vector are shown in Figure 4 and Figure 5, and the results are normalized by the maximum scattering light intensity of the A cement sample with P light incident. It can be seen from the figure that, with the increase of roughness, the scattered light intensity is gradually broadened outside the incident plane. When P light is incident, the broadening rate is higher than that of S light.The Stokes vector V component is almost zero, indicating that the scattered light contains almost no circularly polarized component. At the same time, with the increase of roughness, the difference between Q component and I component becomes larger, indicating that the degree of depolarization of scattered light increases. When P light is incident, the U component has an anti-symmetric peak with respect to the peak position outside the incident plane.When S light is incident, U component has an anti-symmetric peak with respect to peak position in the incident plane.The movement direction of the peak position of Q and U components is basically consistent with that of I components. The scattered light intensity distribution characteristics can be obtained from Stokes vector I component as shown in Table 1. With the increase of roughness, the intensity of scattered light attenuates and the attenuation rate gradually slows down when the incident P light and incident S light occur. The intensity distribution of scattered light is gradually broadened, and the amplitude of incident S light is higher than that of P light. The scattering light intensity peak position moves, P light incident position moves to the direction less than the scattering angle, and S light incident position moves to the position greater than the scattering angle. The shift amplitude of the incident peak position of S light is larger than that of P light. When S light is incident, the intensity of scattered light is higher than P light, which is about 2 times. The angle difference at 0.5 times of the scattered light intensity is used to characterize its width in the incident plane. Table 1Scattering intensity distribution of dielectric rough surface
3.2DiscussionIn order to study the position movement of the scattered light intensity peak, the scattering in the incident plane is selected for analysis. In order to study the position movement of the scattered light intensity peak, the scattering in the incident plane is selected for analysis. As shown in Figure 6, when the incident angle of the local ground element is θ=(θi+θs)/2, then the surface element makes the greatest contribution to the scattering field. At this point, the slope of the surface element is k = tan α=tan[(θi – θs)/2], which satisfies the normal distribution, and its probability density function is , where the average slope μs=0 and the root mean square of the slope . As shown in Figure. 6, when the scattering angle varies from 0° to 90°, the reflectivity of the metal surface is significantly higher than that of the dielectric. The metal surface has little difference in the reflectivity of S light and P light, while the dielectric surface has a large difference in the reflectivity of S light and P light. The product of the reflectivity and the slope probability density function of the microplane element is shown in Figure 7, which is directly proportional to the scattered light intensity.It can be seen that the scattering intensity on the rough surface S light is greater than that of P light, which is about 2 times.When P light is incident, the position of the scattered light intensity peak shifts to the left. When S light is incident, the peak position moves to the right. The scattered light intensity of the metal rough surface is much higher than that of the dielectric rough surface, which is about 20 times that of the incident dielectric rough surface, while the scattered light intensity of the metal rough surface S is slightly higher than that of the P light, but the difference is not significant. P light incident on the rough surface of the metal, the position of the scattered light peak hardly moves. When S light is incident on the rough surface of metal, the position of the scattered light peak shifts to the right, but the shift amplitude is less than that of S light incident on the rough surface of dielectric. 4.CONCLUSIONThe Stokes vector is used to describe the intensity and polarization of the scattered light, and the modulation effect of the incident light P light and S light in two completely polarized states of the cement road surface with different roughness degree is studied. The Stokes vectors of scattered light was measured in hemispheric space. The results showed that the distribution of scattered light intensity is influenced by the reflectivity and the slope distribution of surface element. When S light is incident, the scattered light intensity of the rough surface of dielectric is about 2 times as much as when P light is incident. When S light is incident on the rough surface of metal, the scattered light intensity differs little from P light, and is much higher than the scattered light intensity of dielectric rough surface. The location of the scattered light peak appears different degrees of migration. The rough surface of metal has better polarization-preserving properties. The V component of the scattered light from the cement road surface of dielectric is almost 0. The above studies show that the scattered light has different characteristic distribution with different roughness, so the research on road surface identification can be carried out. It has application value for automatic driving technology road surface recognition. REFERENCESZHANG, H. B.,
“Research on automatic driving technology of vehicle adapted to the condition of curved ramp in underground parking garage,”
Zhenjiang: Jiangsu University, 12
–16
(2011). Google Scholar
“Autonomous driving technology will be included in new car safety assessments,”
Agricultural Equipment & Vehicle Engineering, 56
(04), 16
(2018). Google Scholar
WANG, W. Y.,
“Research on self-tracking of driverless vehicle based on lidar,”
Xi’an: Chang’an University, 14
–16
(2018). Google Scholar
WANG, C. K.,
“Review of autonomous obstacle avoidance technology applied to autonomous driving,”
Chinese Equipment Engineering, 9
(2018). Google Scholar
ZHENG, X. D.,
“Research on intelligent driving automatic control strategy under special roads,”
Chongqing: Chongqing Jiaotong University, 22
–23
(2018). Google Scholar
LIU, Q.,
“The study of electromagnetic scattering method to the ground and sea rough surface,”
Nanjing: Southeast University, 12
–14
(2015). Google Scholar
ZHANG, Y. Y.,
“Study on the electromagnetic scattering characteristics and engineering statistical model of typical land surface,”
Xi’an: Xidian University, 21
–26
(2017). Google Scholar
ZHU, X. M., REN, X. C.,
“Study on electromagnetic scattering from actual road surface with 1D band limited Weierstrass fractal characteristics using FDTD,”
Remote Sensing Technology And Application, 027
(003), 359
–365
(2012). Google Scholar
CAO, Z.,
“Research on electromagnetic scattering from the target above road surface,”
Xi’an: Xidian University, 25
–26
(2007). Google Scholar
LI, E. S., Sarabandi, K.,
“Low grazing incidence millimeter-wave scattering models and measurements for various road surfaces,”
IEEE Transactions on Antennas & Propagation, 47
(5), 851
–861
(1999). https://doi.org/10.1109/8.774140 Google Scholar
Sarabandi, K., ERIC, S.,
“Microstrip ring resonator for soil moisture measurements,”
IEEE Transactions on Geoscience & Remote Sensing,
(1997). https://doi.org/10.1109/36.628789 Google Scholar
Sarabandi, K., LI, E. S.,
“Nashashibi A. Modeling and measurements of scattering from road surfaces at millimeter-wave frequencies,”
IEEE Trans Antennas Propagat, 45
(11), 1679
–1688
(1997). https://doi.org/10.1109/8.650080 Google Scholar
Sarabandi, K., LI, E. S.,
“Polarimetric characterization of debris and faults in the highway environment at millimeter-wave frequencies,”
IEEE Trans Antennas Propagat, 48
(11), 1756
–1768
(2000). https://doi.org/10.1109/8.900234 Google Scholar
|