Research on attitude correction algorithm for mobile wind lidars

The laser wind measurement technology is remarkable for detecting clear-sky wind fields. The Doppler beam swinging algorithm for wind lidars has been developed to obtain vertical wind profiles based on fixed observation methods. However, the Doppler frequencies are superposed due to the self-motions of lidars caused by carrier motions when lidars are used on motion carrier platforms. Meanwhile, the emission directions of laser beams are uncertain due to changes in carriers’ motion directions and tilts. Thus, a new wind measurement correction model must be studied with lidar attitudes. This study considers the influences of the motion velocities, the carrier’s tilt angles, and the laser beams’ yaw angles at the 0° azimuth angle on the measured results under lidar motions, a correction model of motion attitudes for mobile wind lidars was designed. Sensitivity simulation tests for motion attitude parameters were carried out, and the influences of different attitude parameters of the carrier on the measured results were investigated to evaluate and verify the effects of the correction model. Results indicated that the wind measurement correction model could correct data errors caused by the carrier’s motion and tilts. The motion velocities, carrier directions, and the yaw angles of the laser beams at the 0° azimuth angle had an essential influence on the wind velocity measurements. Besides, the carrier’s pitch angles and the roll angles, which did not influence the wind velocity measurements, only affected the altitudes of the wind field data. Furthermore, the pitch angles exerted more significant influences on the data altitudes than the roll angles.


Introduction
Wind lidar is an optical remote sensing technology based on the Doppler principle to detect the velocity of aerosol particles in the atmosphere and then retrieve the atmospheric wind field.The wind lidar [1][2][3][4] has apparent advantages in detection range, data accuracy, spatiotemporal resolution, etc.The vehicle-borne [5,6], ship-borne [7][8][9][10], and airborne wind Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.[11] lidars have been developed based on laser wind measurement technology.They have crucial applications in various areas, including marine environmental monitoring and emergency service guarantee [12,13].Unlike wind lidars in the fixed mode of use, wind lidars on the motion carrier will find the Doppler frequencies superposed due to the self-motions of lidars caused by carrier motions.Moreover, the emission directions of the laser beams are uncertain due to changes in the carrier's motion directions and tilts.Therefore, developing a new wind measurement correction model that considers lidar attitudes is necessary.The wind measurement correction model for vehicle-borne wind lidars was mainly investigated in this study.
There are two main respects for measurement data errors caused by the motions of carrier platforms for wind lidars [14].One is that the motion velocity of the carrier is superposed on the radial velocity measured by the lidar, namely, the motion velocity of air particles relative to the carrier rather than relative to the static object on the ground.The other is that the elevation angle of the beam measured by the lidar servo mechanism when the carrier tilt relative to the ground is the elevation angle of the beam relative to the plane in which the carrier is located rather than relative to the horizontal plane.As a result, there are errors in the altitudes of data in each range bin based on the calculation of the elevation angles of the beams.Two methods are used to solve the influences of carrier motions on data.One is to design a carrier with excellent performance or to equip the carrier with a mechanical compensation device [15].Thus, the carrier can perform the motion compensation, and the influences of the carrier motions on lidar data are minimized as far as possible.The other is to study the motion compensation algorithm, eliminating the attitude influences by the mathematical models [9].The carrier platform is generally equipped with an attitude sensor, given the great difficulty and high costs of the first method.The mathematical attitude correction model is subsequently established, and the wind measurement correction model is studied.
The hardware modification and the compensation algorithm of the wind measurement correction method for mobile wind lidars have been developed.Computer simulations [16][17][18][19] and experimental measurements are both used to test the accuracy of the corrected lidars [20,21].China's first vehicle-borne incoherent wind lidar was developed by the Ocean University of China in 2007 [5].This lidar was installed in a small car, improving the car's maneuverability.However, when the lidar was used to measure data, the car needed to remain stationary.So, it was a kind of fixed mode of use.In 2013, the Ocean University of China designed a damping and vibration isolation system for the carrier platform [22], and the motion tests were also carried out.This system effectively improved the accuracy of the wind measurement under lidar motions.A ship-borne wind lidar was developed, and a motion compensation algorithm was designed by the Ocean University of China in 2015 [23,24].The wind field measurement test of Dongfanghong 2 showed apparent changes in the wind direction during the fixed-point measurement.In the meantime, the wind velocity measured downwind is lower while that measured upwind is higher, so the algorithm needs to be improved.In 2014, a wind lidar using the ocean buoy as a carrier was designed by Delft University, which was equipped on the ocean buoy through the Stewart mechanism [25].The sensor was used to detect the motion attitudes of the ocean buoy, and the motion compensation was carried out so that the lidar could remain horizontal all the time.This method required a high power supply capacity of the ocean buoy at a high price and with a high maintenance cost.Wolken-Mohlmann et al [10] conducted research on the motion compensation theory for ship-borne lidars in Germany in 2014.Besides, the motion compensation algorithm was designed, and tests were carried out to verify the feasibility of the algorithm being applied in offshore wind farms.The algorithm was employed in ocean buoys as well.In 2016, at the University of Tokyo, Professor Yamaguchi and Ishihara [26] summarized a motion compensation algorithm for vertical wind velocity and turbulence intensity measured by lidars through the studies on the influences of carrier motions on the accuracy of wind measurements lidars.The correctness of the algorithm was also verified through simulation tests.Gottschall et al [20] divided the motions of floating lidars into four groups, and analyzed their respective impact on the floating lidar measurement performance.However, the impacts of lidar motion attitudes on data accuracy were not quantified by experiments.In addition, the algorithms proposed by Kelberlau et al [27], Songhua et al [13], Achtert et al [7] and Wolken-Möhlmann et al [10] all used coordinate transformation method to correct the measurement error, which has good implications for us.Kelberlau and Mann [28] used floating lidar system to quantify the influence of amplitude and frequency of motion, the angle between motion and wind direction, and wind speed and strength of wind shear on mean bias with the help of simulation.However, their work needs further refinement, especially needs to add the consideration of motion-measurement on land.
As a result, the mainstream approach to correcting mobile wind lidar data is mechanical compensation based on hardware modification, where only the attitude correction effect of the hardware system on the carrier platform needs to be considered and there is no need to correct the numerical values of lidar data themselves.However, the hardware system has a more complex design and a higher economic cost with a relatively simple principle.Thus, the wind measurement correction algorithm of the motion attitudes for vehicle-borne wind lidars was developed in this study.Although several motion compensation algorithms have been developed, they are mainly applied to the motions of offshore carriers, and the effects of data correction still need to be further researched.In addition, the impacts of lidar motion attitudes on data accuracy were almost not quantified in current study, which could provide important guiding significances for the improvement of mechanical compensation devices.Therefore, we design one attitude correction algorithm for mobile wind lidars based on previous studies, and the influences of different attitude information of the carrier on the measured data were evaluated by the sensitivity simulation tests, verifying the rationality and validity of the model to some extent.The structure of this paper is as follows.The research methods of this paper are mainly introduced in the second chapter, including the overall technical route and the wind measurement correction model.The test designs and results analysis are introduced in the third chapter.The final chapter is about the results and the discussions.

Technical route
The technical route studied in this paper is shown in the figure 1.First, the error sources of mobile wind lidars during the motion measurements, which mainly contained the measurement errors of the radial wind velocities caused by carrier motions and the altitudes of data caused by carrier tilts, were analyzed in this paper.Then, a wind measurement correction model of motion attitudes for mobile wind lidars was designed based on the information on the motion and the attitude parameters of the carrier, aiming at the error sources.The errors in the wind velocities and the data altitudes were corrected one by one.Finally, the model's validity was verified through the sensitivity tests of data simulation.Besides, the influences of various motion and attitude parameters of the carrier on the measured results were compared and analyzed to provide technical support for the later design of the mobile wind lidar system.

Correction model 2.2.1. Model parameters and coordinate system.
Based mainly on the motion parameters, the carrier's attitude information, and the lidar's measured data, the correction model corrected the measurement errors of the radial wind velocities and the data altitudes one by one.There were methods for data correction.For the velocity errors, the carrier motion velocities were projected to the radial velocity directions of the wind lidar.The calculations based on the vector synthesis were carried out so that the influences of the carrier motion velocities on the radial velocities of the lidar were eliminated.For the measurement errors of the data altitudes, the location information of data on the carrier coordinate system was projected into the geographic coordinate system by the coordinate transformations.The geographical coordinate system referred to a right-handed coordinate system in which the coordinate origin was located at the carrier's center of gravity, the positive direction of the X-axis pointed to the due east, while that of the Y-axis pointed to the due north, and the positive direction of the Z-axis was vertical upward.The carrier coordinate system was also a right-handed coordinate system.The coordinate origin was located at the carrier's center of gravity, the positive direction of its X-axis pointed to the front of the carrier, while that of its Y-axis pointed to the left side of the carrier's front, and the Z-axis was vertical to the plane, where the carrier was located, upward.The carrier's course and attitude were the cardinal direction relationship between the carrier and the geographic coordinate systems.
The motion parameters of the carrier in the model were as follows: the motion velocity and the motion direction of the carrier, and the angular velocity due to carrier tilt.The attitude information of the carrier included the pitch and the roll angles of the carrier, and the yaw angle of the laser beam at the 0 • azimuth angle.The elevation angle of the beam, the azimuth angle, the altitudes of data in each range bin, and the radial wind velocity were contained in the measured data of the lidar.Among them, the motion direction of the carrier meant the yaw angle for the forward direction of the carrier, with the 0 • yaw angle in the due north and the clockwise direction as the positive direction.The angular velocity of the carrier represented the angular velocity of the carrier caused by carrier tilt.The pitch angle of the carrier denoted the included angle between the carrier and the horizontal plane when the carrier tilted in the anteroposterior direction.The angle was recorded as a positive number when the carrier tilted upward right ahead.However, when the carrier tilted downward, the angle was recorded as a negative number.The roll angle of the carrier referred to the included angle between the carrier and the horizontal plane when the carrier tilted in the left-right direction.When the carrier tilted upward on the left, the angle was recorded as a positive number.The angle was recorded as a negative number when the carrier tilted downward.The yaw angle of the laser beam at the 0 • azimuth angle was the yaw angle of the laser beam at the 0 • azimuth angle relative to the due north, with the 0 • yaw angle in the due north and the clockwise direction as the positive direction.

Correction of radial wind velocity influences caused by carrier motions.
As for the wind lidar on a mobile platform, the lidar beam angle remains stable when the carrier is translating linearly, so an ideal result of wind measurement is obtained by compensating one velocity vector of the carrier motion for the wind field velocity vector.However, when the carrier is moving, it will inevitably deviate from the motion state of the linear translation under the influences of external factors, such as the turning and the bumpiness of vehicles.Thus the wind lidar detection results may be affected.This study divided the carrier motion into translation and rotation based on different carrier motion conditions.The error sources under the two motion conditions were analyzed respectively, and the motion compensation measures were taken correspondingly.
Firstly, only the carrier translation was considered.When the carrier moved linearly, the line-of-sight vector of wind measurement was merely supplemented with a translational velocity vector, marked as V s , and the projection of the translational velocity vector to the lidar radial direction was symbolized as V sr .The influence component of the carrier motion velocity on the lidar radial wind velocity was obtained, as shown in figure 2.
− → OA is the lidar radial velocity vector, recorded as V w , and − → OB, marked as V ′ w , represents the projection component of the lidar radial velocity vector on the carrier's plane.CE stands for the vertical line of OB on the carrier plane, which is expressed as CE⊥OB, with the vertical foot E. EF is the vertical line of AO on the plane of ABO, shown as EF⊥AO, with the point F as the vertical foot.α and θ represent the elevation angle of the lidar beam of − → OA relative to the carrier-located plane, and the azimuth angle difference between the lidar beam of − → OA and the carrier's motion direction, respectively.
From figure 2, the influence component of the carrier motion velocity on the lidar radial wind velocity was defined as V sr = − → OF, which was testified as below: If the elevation angle of the lidar radial velocity vector (relative to the carrier-located plane) was α, the azimuth angle (relative to the due north direction) was β, and the azimuth angle of the carrier motion velocity vector (relative to the due north direction) was γ.The equations are as below: Then, the corrected lidar radial velocity vector, namely, V correct w , was calculated as: When considering the influence of carrier rotation on lidar, including changes of the pitch angle, the linear velocity direction of the lidar, which was corresponded to the carrier rotational angular velocity, always kept vertical to the radial direction of the lidar beam, so the radial velocity measured by the lidar was not affected.Therefore, only the carrier translation could cause the measurement error of the radial velocity (figure 3 below).

Correction of the altitude deviation corresponding to the radial wind velocity caused by the carrier tilt.
The elevation angle data of the laser beam sent back by the lidar is about the elevation angle of the laser beam relative to the carrier plane, while the azimuth angle data sent back is about the angle of the laser beam relative to the 0 • -azimuth angle direction.The actual elevation angle of the laser beam relative to the ground does not agree with the elevation angle data sent back by the lidar due to the carrier tilt, thus causing the calculation error of the altitude data in all range bins of the lidar.In addition, the actual elevation angle of the lidar relative to the horizontal plane is related to the carrier's tilt angle (the pitch angle and the roll angle were included) and the lidar beam's azimuth angle.As a result, it is challenging to calculate the error angle (the difference between the actual elevation angle and the elevation  angle data sent back by the lidar) directly.To solve the problems, the relative position coordinates in the carrier coordinate system were transformed into the position coordinates in the geographic coordinate system based on the coordinate transformation matrix in this paper.Then, the direct calculation of the error angle was avoided by calculating and synthesizing in the geographic coordinate system.
Then, the coordinate transformation matrix was derived as follows: name two coordinate systems as O − X 1 Y 1 Z 1 and O − X 2 Y 2 Z 2 , and set the former to rotate around the OZ1-axis for angle α to get the latter.Then, let the space vector ⇀ r be projected to the two mentioned coordinate systems to get the projection (x 1 , y 1 , z 1 ) and the projection (x 2 , y 2 , z 2 ), respectively, as shown in figure 4.
The following equation can be obtained from figure 4: Let then r 2 = c 1 r 1 .This equation described the transformation relationship of the same vector projection in different coordinate systems.Specifically, r 1 and r 2 mean the coordinate axes of a particular coordinate vector in the coordinate system 1 and 2, respectively, and c 1 represents the transformation matrix from coordinate system 1 to coordinate system 2. It should be noted that c 1 was just the coordinate transformation matrix rotating around the Z-axis.Similarly, the other two coordinate transformation matrices rotating around the X-axis and the Y-axis were marked as c 2 and c 3 , respectively, to obtain the equations below: Any complicated angular position relationships between the two coordinate systems could be regarded as a limited basic rotational recombination (the angles were set as positive and negative in the counterclockwise rotation and the clockwise rotation, respectively).Therefore, according to the basic rotation order, the transformation matrix equaled the continuous multiplication of the basic rotation-based transformation matrices in the sequence from right to left.Thus, let the laser beam elevation and azimuth angles be φ and θ, respectively.The distance between the data point and the lidar was supposed to be Dis.The carrier's fore-and-aft pitch angle, and the roll angle were assumed as α and β, respectively.Furthermore, assuming the yaw angle of the laser beam at the 0 • azimuth angle to be γ, the azimuth angle in the forward direction of the carrier to be µ, the data point's coordinate in the carrier's coordinate system was expressed as (X, Y, Z): The schematic representation of these angles was presented in figure 5, where − → OA represents the motion direction of carrier on the slope plane where the carrier lies, A ′ is the projection of A on the horizontal plane where the carrier lies, − → OB represents the direction of the lidar radial velocity vector, B ′ is the projection of B on the slope plane where the carrier lies, B ′ ′ As shown below, the transformation matrix c was obtained by the transformation from the lidar coordinate system to the geographic one in the sequence of three steps to avoid influences on the coordinate transformation: (1) rotating around the X-axis to eliminate the influences of the roll angle, (2) rotating around the Y-axis to remove the impacts of the pitch angle, and (3) rotating around the Z-axis to fight off the effects of the yaw angle in the forward direction of the carrier Then, the coordinate position (X 1 , Y 1 , Z 1 ) of the data point in the geographic coordinate system was achieved as follows:

Test design
As wind fields and carrier motion velocity may have impacts on the measurement error, continuous adjustments were made for the carrier's velocity, motion azimuth angle, pitch angle, roll angle, and the yaw angle of the laser beam at the 0 • azimuth angle, based on the wind field of different radial velocities and the constructed wind measurement correction model of motion attitude.Moreover, the influences of different carrier motion parameters on the measured results were evaluated under the background of various radial wind velocities.Besides, the rationality and the validity of the wind measurement correction model of motion attitude were verified.It should be noted that when the sensitivity test was conducted for the motion parameters of a targeted carrier, other carriers' motion parameters were kept constant.The measured result defined in this paper was compared with the actual result to get the error calculation equation as follows: In which, V Error represents the error of the measured result relative to the actual one, N means the number of data, V i stands for the ith actual wind velocity data after correction, and V ′ i symbolizes the ith wind velocity data before correction.

Analysis of the influences of the carrier motion velocity on the wind velocity measurement
Let the azimuth angle of the carrier motion be 20 • , the carrier pitch angle be 10 • , the roll angle be 15 • , the yaw angle of the laser beam at the 0 • azimuth angle be 25 • , the azimuth angle of the laser beam be 60 • , and the laser beam's elevation angle be 75 • .Then, the degree to which the carrier motion velocity affected the measured results was calculated when the radial velocity wind field was under four wind velocity ranges, namely [−1, 1], [−5, 5], [−10, 10], and [−20, 20] (figure 6).Simulation test results under the radial wind    20,20] are respectively illustrated in figures A-D below.The conclusion accorded with the prediction that whatever the background wind velocity is, a smaller carrier velocity indicates a smaller error in the measured result.The root mean squared error (RMSE) of the corrected radial wind velocity relative to the initial radial velocity of the lidar was calculated under the conditions of different background wind velocities and carrier velocities, as shown in table 1.Similarly, regardless of the background wind field, the measurement error, which is irrelative to the background wind velocity, is smaller with a smaller carrier velocity.Meanwhile, the carrier velocity is linearly proportional to the error.This conclusion further verified the validity of the correction model and quantified the effects of the carrier velocity on errors.

Analysis of the influences of the carrier pitch angle on the wind velocity measurement
Let the carrier's motion velocity be 30 km h −1 , and the azimuth angle, the roll angle, the yaw angle of the laser beam at the 0 • azimuth angle, the azimuth angle of the laser beam, and the laser beam elevation angle be 10   As listed in table 2, the RMSE of the corrected wind velocity relative to the initial velocity was calculated with different background wind velocities and carrier pitch angles.The RMSE of the corrected wind velocity relative to the initial velocity is always 3.2814, regardless of the background wind velocity and the carrier pitch angle.Therefore, it could be concluded that there is no impact of the carrier pitch angle on the wind velocity measurement, irrespective of the background wind field.The RMSE value, 3.2814, is caused by other carrier motion parameters.

Analysis of the influences of the carrier roll angle on the wind velocity measurement
With the motion velocity of the carrier set at 30 km h −1 , the azimuth angle of the moving carrier of 10 • , the pitch angle of 15 • , the yaw angle of the laser beam at the 0 • azimuth angle of 25 • , the azimuth angle of the laser beam of 60 • , and the elevation angle of the laser beam of 75 • , the degree to which the different roll angles of the carrier affected the wind velocity measurement was calculated when the radial wind velocity was set within four wind velocity ranges: [−1, 1], [−5, 5],     Table 4. RMSE of the corrected wind velocity relative to the initial velocity under the conditions of different background wind fields and carrier motion directions.Table 3 shows the RMSE of the corrected radial wind velocity relative to the initial radar radial velocity calculated under the conditions of different background wind velocities and carrier roll angles.The RMSE of the corrected wind velocities      20,20], respectively.The results suggest that a smaller azimuth angle of the carrier motion results in a smaller RMSE in the wind velocity measurement, regardless of the background wind velocity.Table 4 shows the RMSE of the corrected radial wind velocity relative to the radar radial velocity calculated under the conditions of different background wind velocities and motion directions of the carrier.The magnitude of the background wind velocity makes little difference to the measurement error at the same azimuth angle of the carrier motion.Meanwhile, a positive linear relationship exists between the azimuth angle of the carrier motion and the measured RMSE.

Analysis of the influences of the yaw angle of the laser beam at the 0 • azimuth angle on wind velocity measurement
With the motion velocity of the carrier set at 30 km h −1 , the azimuth angle of the carrier motion of 20 • , the pitch angle of 10 • , the roll angle of 15 • , the azimuth angle of the laser beam of 60 • , and the elevation angle of the laser beam of 75 • , the degree to which the different yaw angles of the laser beam at the 0 • azimuth angle on the wind velocity measurement was calculated when the radial wind velocity was set within four wind velocity ranges:  20,20], respectively.The results suggest that the measurement error will be smaller for the larger yaw angle of the laser beam at the 0 • azimuth angle, regardless of the background wind velocity, which is contrary to the influences of the azimuth angle of the carrier motion on wind velocity measurement.
Table 5 shows the RMSE of the corrected radial wind velocity relative to the initial radar radial velocity calculated under the conditions of different background wind velocities and yaw angles of the laser beam at the 0 • azimuth angle.The measurement error appeared to be unaffected by the background wind velocity under the same yaw angle of the laser beam at the 0 • azimuth angle.Meanwhile, the yaw angles of the laser beam at the 0 • azimuth angle are inversely proportional to the linearly measured RMSE.

Analysis of influences of the carrier pitch angle and the roll angle on the altitude of data
The tilt of the carrier, including the pitch and roll angles, contributes a lot to the altitude error of data.With the carrier motion velocity set at 30 km h −1 , the motion azimuth angle of 20 • , the yaw angle of the laser beam at the 0 • azimuth angle of 25 • , the azimuth angle of the laser beam of 60 • , and the elevation angle of the laser beam of 75 • , the altitude of data calculated under the conditions of different pitch angles and roll angles are shown in figure 11.Smaller carrier pitch and roll angles suggest a smaller influence on the altitude of the data.Compared with the roll angle, the carrier pitch angle conferred a greater magnitude of influence on the altitude of data.

Test summary
To further compare the influences of each motion parameter of the carrier on the measurement results, search out the factors that exert the most significant impact on the measurement results, and avoid errors through hardware design, the influences of unit carrier motion velocity, unit carrier tilt angle, unit azimuth angle of the carrier motion and unit yaw angle of the laser beam at the 0 • azimuth angle on the measured radial wind velocity were calculated with the results shown in table 6.The wind velocity measurement is significantly affected by the carrier motion velocity, carrier motion direction, and the yaw angle of the laser beam at the 0 • azimuth angle.The yaw angle presents the most significant impact, followed by the carrier motion azimuth angle and the carrier motion velocity.Besides, the carrier's pitch and the roll angles do not affect the wind velocity measurement.However, they only affect the altitude of data.The pitch angle exerts more influence on the altitude of data than the roll angle.

Conclusion and discussion
The effect mechanism of carrier motions on measurement results was analyzed to apply the mobile wind lidars on vehicles, and a motion attitude correction algorithm was designed in this study based on carrier attitude sensor information.Besides, based on sensitivity simulation experiments, the degree to which the different carrier motion parameters on measurement results were calculated and the rationality and validity of the motion attitude correction algorithm proposed in this paper were verified.Finally, the following conclusions can be drawn: (1) Wind velocity measurement is significantly affected by the carrier motion velocity, carrier motion direction, and the yaw angle of the laser beam at the 0 • azimuth angle, among which the yaw angle presents the most significant impact, followed by the carrier motion azimuth angle and the carrier motion velocity.(2) The carrier pitch and the roll angles do not affect wind velocity measurement.However, they only affect the altitude of data.The pitch angle exerts more influence on the altitude of data than the roll angle.
The study results are mainly used to correct the radial wind velocity data under the wind lidar motion conditions.The three-dimensional wind field in the observation area can be retrieved by combining the corrected radial wind velocity data and the Doppler beam swinging/velocity azimuth display wind retrieval algorithm.The research results will lay essential technical support for the development of mobile wind lidar hardware systems with relatively high practicability.However, it should be noted that the accumulation effect of lidar cannot be ignored when we used it on one unstable platform, which will affect the quality of inversion data and need further studies, although the issue is not the main focus of this paper.In addition, since the algorithm proposed here is verified only based on numerical simulations, further research is required to confirm its practical effect.Therefore, outdoor mobile observation tests should be carried out in the future, and the wind data corrected by the algorithm and the results of other fixed wind measurement methods should be compared and analyzed further verify the effectiveness of this proposed algorithm.Actually, we are using lidar to conduct the experiments to detect low-level wind shear in the airport with our previously proposed algorithm, which was also only verified by numerical simulations before.Outdoor mobile observation experiments with wind lidars will be carried out once the low-level wind shear detection experiments were completed.

0Figure 1 .
Figure 1.The technical route of this paper.

Figure 2 .
Figure 2. Correction model of the translational carrier wind field vector.

Figure 3 .
Figure 3. Correction model of the rotational carrier wind field vector.

Figure 4 .
Figure 4. Rotation of the coordinate system.

Figure 5 .
Figure 5.The schematic representation of different angles, including the laser beam elevation angle, azimuth angle, the carrier's fore-and-aft pitch angle, the roll angle, the yaw angle of the laser beam at the 0 • azimuth angle, and the azimuth angle in the forward direction of the carrier.

Figure 6 .
Figure 6.Comparison of the influences of the carrier velocities on the measured results in different wind velocity ranges.

Figure 7 .
Figure 7.Comparison of the influences of the carrier pitch angle on the measured results in different wind velocity ranges.

Figure 8 .Table 3 .
Figure 8.Comparison of the influences of the carrier roll angle on measurement results in different wind velocity ranges.

Figure 9 .
Figure 9.Comparison of the influences of the carrier motion directions on measurement results in different wind velocity ranges.

Figure 10 .
Figure 10.Comparison of the influences of the yaw angle of the laser beam at the 0 • azimuth angle on measurement results in different wind velocity ranges.

Table 5 .
RMSE of the corrected wind velocity relative to the initial velocity under the conditions of different background wind fields and yaw angles of the laser beam at the 0 • azimuth angle.Background wind velocity Yaw angle of the laser beam at the 0 • azimuth angle [−1 m s −1 , 1 m s −1 ] [−5 m s −1 , 5 m s −1 ] [−10 m s −1 , 10 m s −1 ] [−20 m s −1 , 20 m s −1 ] 10