Design of a Liquid-Driven Laser Scanner with Low Voltage Based on Liquid-Infused Membrane

This study proposes a liquid-driven laser scanner installed on the end effector of a continuum endoscope to perform accurate laser steering in a constrained environment. In contrast with previous studies, the ﬁ rst contribution of this work is using a transparent miniature prism (7 mm (cid:4) 7 mm) based on opto ﬂ ui-dic technology as the laser scanner. The opto ﬂ uidic prism has no mechanical moving component, eliminating the mechanical fatigue of existing mechanical laser scanners and providing a compact structure. The second contribution is an innovative method for fabricating an opto ﬂ uidic prism based on the liquid-infused membrane, demonstrating stable and reliable


Introduction
Laser energy has been used extensively in precise manipulations, such as welding [1] and micromachining [2] in industries and tumor ablation [3] and wound suturing [4] in surgical operations, to name a few.The use of flexible optical fibers to deliver laser energy can avoid the straight working path of rigid optical components. [5,6]9][10] Manipulating the fiber tip is the most straightforward solution for laser steering; this technique includes cable-driven, [7] electromagnetic, [8] and hydraulic [9] methods.Cable-driven laser scanners have simple structures, but they experience difficulty in scanning the target by bending the optical fiber back and forth within narrow cavities.Electromagnetically driven laser scanners have a low driving voltage but suffer from hysteresis.Hydraulically driven laser scanners are immune to external magnetic fields (e.g., magnetic resonance imaging) but have a scanning speed as low as 2 mm s À1 . [9]In contrast, mechanically controlled mirrors driven by piezoelectric motors are used to achieve laser steering. [10]However, the driving voltage of piezoelectric motors is up to 200 V, which is too high for the human body.Moreover, additional components are required to maintain the optical axis for reflective devices, resulting in a bulky size and an obstructed field of view.At present, optofluidic laser scanners that are characterized by a transmissive mode and a nonmechanical structure provide an economical solution for laser steering. [11]ompared with existing mechanical laser scanners, nonmechanical laser steering is achieved with an optofluidic prism based on electrowetting on dielectric (EWOD). [12,13]The optofluidic prism has no mechanical moving component, and its power consumption is down to microwatts. [14,15]The optical axis can be maintained due to the prism's transmissive mode, providing a compact structure and fast operation. [16,17]However, charge trapping in the water-solid interface of an EWOD device can lead to the degradation of the dielectric layer and ultimately, to a breakdown in tens of cycles. [18,19]Although increasing the thickness of the dielectric layer or coating a dielectric material with a large dielectric constant can prevent breakdown, these processes involve high driving voltages, sophisticated manufacturing processes, and degraded electrowetting performance. [20,21]or example, an EWOD device will fail in 100 cycles with a 200 nm-thick Ta 2 O 5 (dielectric constant ε ¼ 23) film due to charge trapping. [22]Therefore, the poor stability of existing EWOD devices needs to be further improved.
[25] On the one hand, such a composite membrane avoids the breakdown of the dielectric layer; on the other hand, it facilitates trapping charge migration.For example, droplets driven by EWOD have achieved more than 5000 repeatable actuations on a liquid-infused membrane. [19]However, the maximum driving voltage of EWOD devices based on the liquid-infused membrane can reach as high as 400-500 V.An excessively high driving voltage raises challenges for practical applications of liquid-infused membranes to EWOD devices, in addition to laser-assisted surgery.
This study proposes a liquid-driven laser scanner installed on the end effector of a continuum endoscope to perform accurate laser steering in a constrained environment.In contrast with previous studies, the first contribution of this work is using a transparent miniature prism (7 mm Â 7 mm) based on optofluidic technology as the laser scanner.The optofluidic prism has no mechanical moving component, eliminating the mechanical fatigue of existing mechanical laser scanners and providing a compact structure.The second contribution is an innovative method for fabricating an optofluidic prism based on the liquid-infused membrane, demonstrating stable and reliable laser steering in 1000 cycles.The third contribution is that the maximum driving voltage of the liquid-infused membrane-based laser scanner is reduced to 15 V, ensuring safety in biomedical applications.Finally, the laser steering experiment agrees with the established theoretical model, verifying the laser scanner's controllability and reliability.

System Overview
The designed laser steering device is presented in Figure 1.When using a continuum robot to project laser energy, operators first search the target site by controlling the flexible endoscope and then define the ablation area. [26]Thereafter, the laser spot will be steered automatically with the developed liquiddriven laser scanner to perform laser ablation, as shown in Figure 1a.The endoscope had a length of 1 m, and a fiber collimator was inserted into the endoscope's working channel to collimate the fiber-transmitted laser from the laser source, as illustrated in Figure 1b.Then, the optofluidic prism, which works as the laser scanner, was affixed onto the endoscope at a height of z (80 mm) for testing.A laser with a 650 nm wavelength was emitted from the fiber collimator to the scanner.A camera was used to monitor the operating status of the optofluidic scanner, while a computer was used to send the control signal to the microcontroller (Arduino MEGA2560).Subsequently, the pulse width modulation (PWM) signal generated by the microcontroller was sent to the driver module (L298N) to develop a 1 kHz AC square wave to actuate the scanner.The amplitude of the square wave was programmed by adjusting the duty cycle of the PWM signal, manipulating the laser steering angle of the laser scanner.
The current study used an optofluidic prism as the liquiddriven laser scanner.As shown in Figure 1c, the fabricated optofluidic prism had a diameter of 7 mm and a length of 7 mm, and a weight of %1 g; thus, it was sufficiently small to be integrated into the endoscope while maintaining an unobstructed endoscopic view.Figure 1d,e shows the endoscopic view without and with the developed laser scanner, respectively.The visual field remained nearly intact due to the transparent structure of the optofluidic prism.

Design and Fabrication of the Optofluidic Prism
In this section, the design and fabrication of the optofluidic prism were conducted.The optofluidic prism consisted of four components: 1) two immiscible media, of which, water was the most commonly used conductive medium; 2) a dielectric layer that acted as a capacitor to prevent electrolysis; 3) a hydrophobic layer to increase the initial contact angle and further increase contact angle variation; and 4) conductive electrodes to apply voltages.The optofluidic prism was designed as a cuboid to achieve a flat liquid interface.In contrast with cylindrical prisms, which require 16 or 32 electrodes to achieve a nearly flat liquid interface, [27,28] the cuboid prism can realize an utterly flat liquid interface by matching only four contact angles on the four electrodes, undoubtedly simplifying the drive system.In addition, the processing and assembly of the cuboid prism are relatively simple, making it an economical scheme for fabricating the optofluidic prism.
The designed optofluidic prism was assembled with five pieces of indium tin oxide (ITO) glass (Guluo glass, Luoyang, China) as the sidewalls and bottom and one bare glass as the top.The cuboid prism had an inner side length of 3 mm and a height of 5 mm, with an additional 2 mm width of the electrode and 1 mm glass thickness on each side.Figure 2 illustrates the fabrication process of the optofluidic prism.As shown in Figure 2a, four pieces of ITO glass with a size of 5 mm Â 6 mm and a square resistance of 6 Ω were sequentially cleaned with alcohol and deionized water for 10 min, dried with nitrogen, and then cleaned with plasma for 10 min.After cleaning, the ITO glasses were covered along the short sides using a 2 mmwide polyimide tape to preserve ITO electrodes of 2 mm Â 5 mm, as shown in Figure 2b.A tenfold dilution of epoxy resin (Nanjunhui, Shenzhen, China) was then coated onto the ITO glasses to form an %1 μm-thick dielectric layer, as shown in Figure 2c.Thereafter, a 1 wt% Cytop (CTL-809M) was spin coated and baked at 110 °C for 60 min as the %0.1 μm-thick [29] hydrophobic layer to ensure a large contact angle-tuning range.Then, the ITO glasses were assembled with optical epoxy under ultraviolet light.Subsequently, the prism was first filled with silicone oil (5 cSt, Dow Corning) and then inverted for 30 min to form the liquid-infused membrane, as shown in Figure 2e,f.Finally, equal volumes of 0.1 wt% sodium dodecyl sulfate aqueous solution (ρ ¼ 1 g cm À3 , n 2 ¼ 1.33) and oil (ρ ¼ 1 g cm À3 , n 1 ¼ 1.58) were sequentially filled in the prism, wherein oil was a mixture of 1-bromonaphthalene (1.48 g cm À3 , 3.22 cSt, Rhawn, Shanghai, China) and dodecane (0.75 g cm À3 , 1.79 cSt, Rhawn, Shanghai, China), to achieve density matching.After sealing with a bare glass, the optofluidic prism based on the liquid-infused membrane was fabricated, as shown in Figure 1c and 2g.

Working Principle
The designed liquid-driven laser scanner achieves laser steering on the basis of optical refraction.As shown in Figure 3a, the optofluidic prism was filled with immiscible, transparent, and density-matched oil and water with different refractive indices.Two refractions occurred after the incident light passed through the oil-water interface and the water-air interface.The final refraction angle after the two refractions was the laser steering angle.In accordance with Snell's law, the laser steering angle δ (°) can be expressed as where n 0 , n 1 , and n 2 are the refraction indexes of air, oil, and water, respectively; φ (°) is the tilt angle of the oil-water interface.When δ > 0, the outgoing beam was deflected toward the left; when δ < 0, it was deflected toward the right.The optofluidic prism adjusts the laser steering angle by changing the tilt angle of the liquid interface on the basis of EWOD.EWOD balances the interfacial tension of the solidoil-water phases at different contact angles by applying voltages.In accordance with the Young-Lippmann equation, the contact angle can be described as where θ i (°) is the contact angle at voltage U i (V ), i ¼ 1, 2, 3, 4; θ 0 (°) is the initial contact angle; ε is the relative permittivity of the solid dielectric impregnated with lubricant; [30] ε 0 (F m À1 ) is the permittivity of the vacuum; d (m) is the thickness of the composite dielectric; and γ (N m À1 ) is the interfacial tension of the oil-water interface.
The oil-water interface should be straight to ensure the same refraction angle of the beam at any point through the oil-water interface; that is, the two opposing contact angles should satisfy , the laser steering angle δ can be controlled by adjusting the contact angle θ 1 on the basis of Equation ( 2).Therefore, the relationship between the voltages on the electrodes and the position of the laser spot was investigated.
The position of the laser spot on the projection plane can be expressed in polar coordinates as (δ, ψ, z), where the laser steering angle δ and the height z from the prism to the projection plane determine the distance from the center; the rotation angle ψ (°) determines the angular position of the spot.Therefore, the contact angles can be described by the tilt angle φ and the rotation angle ψ [17] 8 > > > < > > > : By combining Equation (1-3), the applied voltages can be expressed as where Equation ( 4) demonstrates the controllability of the laser steering device.Successively, the contact angles were examined at different voltages, based on which the voltage combinations could be determined accordingly.
The optical and fluid simulation was performed using COMSOL software.Figure 3b presents the simulation result with the tilt angle φ ¼ 35°and the rotation angle ψ ¼ 0°.A laser steering angle of 10.6°was achieved in the simulation, which was sufficiently large for biomedical applications. [10]The rapid response of the liquid driven by EWOD is essential for fast laser steering.The simulation results showed that the liquid interface became flat and the tilt angle reached 35°after applying voltages for 200 ms.The white arrows indicate the directions of fluid velocity.The closer the fluid flowed to the liquid interface, the more violent the flow became.However, the tilt angle of the liquid interface, which determines the laser steering angle, remained constant, indicating the dynamic equilibrium of the optofluidic system during operation.

Liquid-Infused Membrane
The liquid-infused membrane, which is formed by the diffusion of the nonvolatile lubricant onto the textured surface of the solid, is a stable lubricant film.The solid surface texture of the optofluidic prism is formed by the interaction of the epoxy resin polymer network of the dielectric layer and the Cytop of the hydrophobic layer. [31,32]The optofluidic prism was first filled with silicone oil (Dow Corning, 5 cSt, surface tension: 19.7 mN m À1 ) and then inverted for 30 min to form the composite liquid-infused membrane.Silicone oil was chosen as the lubricant because of its high dielectric constant (2.65) and nonvolatile nature.In addition, a low-viscosity silicone oil of 5 cSt produces a small viscous resistance and facilitates reversible electrowetting. [19]The inversion time determines the membrane thickness, which affects the electrowetting response and dielectric breakdown. [33]The thickness of the liquid-infused membrane can be estimated by gravimetric analysis. [25]In our experiment, the thickness of the liquid-infused membrane was %60 μm.
After lubricating with silicone oil, the infused silicone oil climbed to the water-filled surface and formed a thin shell that acted as both a lubricating and insulating layer, [34] as shown by the yellow line in Figure 4a.Instead of the direct contact between the conductive liquid and the solid, the lubricant film works as an insulating layer for the lubricant-impregnated surface, avoiding dielectric breakdown. [18]For example, electrolysis occurs and bubbles come out when applying voltages to the optofluidic prism without lubrication, as shown in Figure 4b.In contrast, the conductive liquid was encased in an oil shell when lubricated with silicone oil to form a liquid-infused membrane, as shown in Figure 4c.Electrolysis caused by dielectric breakdown was eliminated, resulting in a perfect electrowetting performance.In addition to reducing friction with the solid, the migration of trapped charges at the water-solid interface into the aqueous solution was facilitated.Consequently, the contact angle hysteresis resulting from the pinning of the oil-watersolid three-phase contact line was eliminated. [25]Therefore, the liquid-infused membrane not only avoids dielectric breakdown but also facilitates trapping charge migration and eliminates contact angle hysteresis.

Contact Angle Measurement
The designed liquid-driven laser scanner adjusts the tilt angle of the liquid interface by matching the contact angle to determine the spot position.To achieve voltage-controlled and programmable laser steering, the relationship between the driving voltage and contact angle of the optofluidic prism was first investigated.As shown in Figure 5, the optofluidic prism was placed straight [(a), (b), (c)], horizontally [(d), (e), (f )], and flipped vertically [(g), (h), (i)].All the initial contact angles in the three statuses were 160°due to the density matching of the filled oil and water.Then, the liquid interfaces in the three states were monitored when applying U 1 and U 2 to the two opposite electrodes.The tilt angle of the liquid interface remained at 35°when regardless of the placement of the optofluidic prism.In addition, due to the density-matching oil and water in the prism, the volume ratio of oil and water will not affect the maximum tilt angle as long as the electrowetting effect functions.These experimental results demonstrate that the laser steering angle is only related to the refractive indexes of the filled liquids and the tilt angle of the liquid interface regardless of the postures of the designed prism.
Figure 6 illustrates the relationship between the applied voltage and the contact angle when the optofluidic prism is placed straight, horizontally, and flipped vertically.The shaded curves indicate the standard deviation.The contact angle was measured using the software ImageJ, and three measurements were taken at each voltage node.For example, the standard deviation of the contact angle at 7 V was lower than 0.4°, indicating the highly repeatable measurement results.The contact angle curves under different postures largely overlaped, and the contact angle gradually decreased with the increase of the voltage.Moreover, the contact angle reached %55°at 15 V, and no more extended changes occurred with continued voltages.This contact angle saturation corresponds to a contact angle variation range of 105°.Notably, the electrowetting threshold voltage of the optofluidic prism based on the liquid-infused membrane is extremely low, such that a significant change occurs in the contact angle at 1 V (145°).

Laser Steering Ability
After determining the relationship between the contact angle and the voltage, the laser steering ability of the liquid-driven laser scanner was tested.As shown in Figure 7, different voltages of U 1 and U 2 were applied to the laser scanner for linear scanning.The liquid interface was horizontal at U 1 ¼ U 2 ¼ 9 V, and no beam deflection occurred.The beam deflected toward the left at U 1 ¼ 15 and U 2 ¼ 2 V, while it deflected toward the right at U 1 ¼ 2 and U 2 ¼ 15 V, both with a liquid interface tilt angle of 35°and a spot displacement of 15.0 mm.The laser spot displacement was set as D, and then the laser steering angle can be described as δ ¼ arctan D/z, from which the maximum laser angle was calculated as 21.2°(AE10.6°).
Further experiments were conducted to achieve continuous 2D laser steering on the projection plane.As shown in Figure 8a, the liquid interface became a horizontal plane at such that the laser beam could directly pass through the optofluidic scanner without refraction.The initial projection of the spot was set as the coordinate origin.Then, programmable voltages were applied to allow the laser spot to achieve a scanning of 360°along the vertical, horizontal, and  diagonal directions.For example, at U 1 ¼ 15, U 2 ¼ 2, and U 3 ¼ U 4 ¼ 9 V, the laser spot was projected onto the negative x-axis, as shown in Figure 8b.Meanwhile, at U 1 ¼ U 4 ¼ 2 and U 2 ¼ U 3 ¼ 15 V, the laser spot was projected onto the diagonal direction, as shown in Figure 8c.Moreover, the tilt angle of the liquid interface in Figure 8b was 27°, which was smaller than the tilt angle when using only two opposite electrodes for linear scanning.The reduction of the tilt angle was due to the oil isolation of the cuboid prism, where oil fills the corners and penetrates into the bottom to interfere with prism operation. [17]igure 8d presents the maximum displacement of the laser spot in eight directions on the projection plane.Figure 8e presents the maximum laser steering angle and standard deviation over three scanning cycles, where the red circles represent the locations of the spot.The maximum laser steering angles in the lateral and longitudinal directions were 13.36°and Locations of the laser spot on the projection plane.e) Maximum laser steering angles over three cycles.
13.25°, respectively.In contrast, the maximum laser steering angle in the diagonal direction could reach 21.32°, with a standard deviation as low as 0.02°.The laser steering angle difference in the axial and diagonal directions resulted from the assembly errors of the optofluidic prism. [35]n addition, the repeatability of the laser scanning paths in the horizontal, vertical, and two diagonal directions was measured, as shown in Figure 9.The repeatability was indicated in terms of the standard deviation of three laser scanning paths in each direction.The locations of the spot were recognized through image processing and expressed in pixels (95 μm pixel À1 ).Table 1 indicates the repeatability and maximum error of the laser scanning paths in different directions.The vertical line has the largest repeatability of 107.8 μm and a maximum error of 440.0 μm, while diagonal line A has the smallest repeatability of 78.9 μm and a maximum error of 264.9 μm.All the repeatability and maximum errors meet the clinical laser targeting accuracy (<1 mm), [36] implying the working reliability of the liquid-driven laser scanner.

Stability of the Laser Scanner
The stability and reliability of the liquid-driven laser scanner are essential for practical applications.EWOD device breakdown occurs in less than 10 cycles with 800 nm-thick Teflon AF 1600X as the dielectric and hydrophobic layers. [19]In this regard, this study investigated the stability of the laser scanner based on the liquid-infused membrane.On the one hand, the liquidinfused membrane acts as an insulating layer to avoid dielectric breakdown and facilitate trapping charge migration.On the other hand, the liquid-infused membrane works as a lubricating layer to eliminate contact angle hysteresis caused by contact line pinning, improving the stability of the optofluidic device.Compared with existing low-voltage-actuated EWOD devices that can only work for tens of cycles, [19,22] the proposed liquiddriven laser scanner can perform 1000 cycles of linear scanning, as shown in Figure 11.The x-axis denotes the cycle numbers, the y-axis denotes the tilt angle of the liquid interface, and the colors represent different laser steering angles, as shown in Figure 11a.The tilt angle changes continuously from the left limit to the right limit and back to the left limit in each cycle.Figure 11b,c shows the statistics of the tilt angle of the left limit and right limit over 1000 cycles.Consequently, no evident degradation of the tilt angle and the laser steering angle was observed over 1000 cycles.Moreover, the maximum laser steering angles over 1000 cycles were counted, wherein the left limit was 10.58°AE 0.30°, while the right limit was À10.57°AE 0.45°.Both standard deviations were within 5%, indicating the excellent stability of the laser scanner.This experiment demonstrates that the designed scanner can achieve stable and reliable laser steering for as long as 1000 cycles.

Response Time and Reflection Loss
The moving speed of the spot determines the amount of energy projected onto the target when using laser energy delivered by a continuum robot.This experiment investigated the response time of the optofluidic prism to evaluate the maximum laser steering speed.The response time is defined by the time it takes the laser to sweep 5 mm from the initial point, given that the average length of an oral tumor is 4.04 mm. [37]A camera with 60 frames s À1 was used to record the video, which took 16.67 ms frame À1 .It took 11 frames to sweep 5 mm either left or right from the initial point, indicating that the optofluidic prism's response time was about 183.3 ms.Hence, the maximum laser steering speed was 27.3 mm s À1 .In addition, reflection loss occurs when the laser passes through the oil-water and water-air interfaces.In accordance with the Fresnel equation, reflection loss R is where R s and R p are the reflectance of s-polarized light and p-polarized light, respectively; θ i is the incident angle and θ t is the refraction angle at each interface. [38]The reflection loss versus tilt angle φ curve was plotted, as shown by the red curve in Figure 10.Reflection loss increased with tilt angle.When the tilt angle of the liquid interface was 35°, the reflection loss was 3.1%.

Discussion and Conclusion
This study proposed a liquid-driven laser scanner installed on the end effector of a continuum endoscope and developed an economical and reliable fabrication method for optofluidic devices.
As indicated in Table 2, compared with existing methods, the proposed method exhibits the advantages of miniature size, low driving voltage, transparent structure, transmissive mode, and high stability.A liquid-infused membrane-based optofluidic prism was designed and experimentally fabricated for precise laser steering.Such a membrane facilitated trapping charge migration and hampered dielectric breakdown, reducing driving voltage from 500 to 15 V.With the low-voltage-actuated laser scanner, the 2D laser scanning paths demonstrated excellent repeatability as low as 78.9 μm, and the maximum laser steering angle presented no evident degradation in 1000 cycles when linear scanning.Moreover, an unobstructed endoscopic view was realized due to the laser scanner's transparent structure and miniature size.Finally, reflection loss and maximum steering speed were measured.Some issues remain to be improved in future research.First, the refractive index difference between oil and water is finite due to the requirement of oil-water density matching in the laser scanner, resulting in a limited laser steering angle.Therefore, an optofluidic scanner with multiple liquid interfaces can be investigated to allow multiple optical refractions to expand the laser steering range.Second, the laser scanner encounters challenges in fabrication with the uneven thickness of the dielectric layer and substandard manual assembly, reducing laser targeting accuracy.In the future, we will improve trajectory tracking accuracy with a learning-based control method to facilitate precise scanning of the target site in laser-assisted surgery.Working principle Zhao et al. [7] 22 22 -Cable-driven Acemoglu et al. [8] 13 60 9 Electromagnetic Fang et al. [9] 12 100 -Hydraulic York et al. [

Figure 1 .
Figure 1.The laser steering device.a) Schematic of the liquid-driven laser scanner on a continuum endoscope.b) Experiment setup of the designed system.c) The optofluidic prism installed on the end effector of the endoscope.d,e) Endoscopic view without and with the developed laser scanner, respectively.

Figure 2 .
Figure 2. Fabrication process of the optofluidic prism.a) Clean four pieces of ITO glasses as the sidewalls.b) Use polyimide tape to preserve ITO electrodes.c) Coat epoxy and Cytop as the dielectric layer and hydrophobic layer, respectively.d) Assemble the ITO glasses.e,f ) Inject silicone oil and then invert for 30 min to form the liquid-infused membrane.g) Schematic of the optofluidic prism after liquid dosing and sealing.U 1 , U 2 , U 3 , and U 4 indicate the driving voltages on the ITO electrodes.

Figure 3 .
Figure 3. Laser steering principle of the optofluidic prism.a) Mechanism of the prism when voltages are applied to two opposite electrodes.b) Simulation when the tilt angle of the liquid interface is 35°.

Figure 5 .
Figure 5. Liquid interfaces of the optofluidic prism under different postures when applying U 1 and U 2 .a-c) Liquid interfaces when the prism is placed straight.d-f ) Liquid interfaces when the prism is placed horizontally.g-i) Liquid interfaces when the prism is flipped vertically.

Figure 6 .
Figure 6.Relationship between the applied voltage and contact angle of the optofluidic prism under different postures.

Figure 7 .
Figure 7. Measurement of the maximum laser steering angle.
With reference to the laser steering experiment, a laser steering angle versus tilt angle of the liquid interface curve was plotted, as shown in Figure 10.The blue curve indicates the theoretical value based on Equation (1), while the red spots indicate the experimental value.The theoretical and experimental values are in good agreement, demonstrating the excellent controllability and stability of the scanner.

Figure 9 .
Figure 9. a) Laser scanning paths in the horizontal and vertical directions, and b) the two diagonal directions.

Figure 11 .
Figure 11.1000 cycles of linear scanning using the liquid-driven laser scanner.a) Relationship between the tilt angle and the laser steering angle over 1000 cycles.b) Histogram of the left limit tilt angle.c) Histogram of the right limit tilt angle.

Table 1 .
Repeatability of laser scanning paths in different directions.
Figure 10.Relationship among reflection loss, laser steering angle, and tilt angle.

Table 2 .
Comparison with state-of-the-art methods.