- 1. Usage
- 2. Calibration Result
1. Usage
1.1 Data Collect Procedure
More than three motions in different positions and orientations are required, so that the equation is full rank and can be solved linearly.
Therefore, the algorithm requires the trajectory needs to contain translations and rotations(like vehicle to travel in the shape of figure 8, as shown in Figure down below).
The point cloud and GNSS+INS data of the trajectory need to be recorded simultaneously.
There should be rich features in the experimental environment, like buildings, and less moving objects, like moving vehicle, which will affect the result of Lidar odometry.
Do not test in an environment surrounded by tall buildings, this will affect the accuracy of GNSS.
Could record more than one sets of data for verification if necessary.
1.2 Precision Indicators
estimate pose:
![]()
reference pose:
![]()
1.2.1 RPE: relative pose error
i frame RPE:
![]()
1.2.2 ATE: absolute trajectory error
1.3 in & out
Input:
1
2
3
4
5
6
// PointCloud
pcl::PointCloud<pointXYZ> point_in;
// IMU pose
std::vector<Eigen::Matrix4d> IMU_pose;
// PointCloud timestamps
std::vector<double> timestamps;
Output:
1
2
3
4
5
6
//Extrinsic Matrix:
Eigen::Matrix4d calibratedTransformation;
/*
| R T |
|_ 0 1 _|
*/
2. Calibration Result
2.0 Extrinsic Matrix
1
2
3
4
[[0.9996387536200673, -0.01085860442221318, 0.02458562529040346, -1.697536587497878],
[-0.01096721316998588, -0.9999306683289505, 0.004287046827647254, -0.007142523421813133],
[0.02453736938227735, -0.004555133940977415, -0.99968853562426, -1.502642065472676],
[0, 0, 0, 1]]
2.1 Calib Data:(0655)
2.1.1 Traj
2.1.2 XYZ
2.1.3 RPY
2.1.4 APE
1
2
3
4
5
6
7
max 0.307479
mean 0.123667
median 0.122317
min 0.009096
rmse 0.133519
sse 19.217960
std 0.050338
2.1.5 RPE
1
2
3
4
5
6
7
max 0.125233
mean 0.031746
median 0.027947
min 0.002622
rmse 0.038173
sse 1.569387
std 0.021198
2.2 Valid Data:(0657)
2.2.1 Traj
2.2.2 XYZ
2.2.3 RPY
2.2.4 APE
1
2
3
4
5
6
7
max 0.352818
mean 0.108570
median 0.095848
min 0.006885
rmse 0.127275
sse 13.347926
std 0.066419
2.2.5 RPE
1
2
3
4
5
6
7
max 0.183633
mean 0.029524
median 0.026837
min 0.001288
rmse 0.034410
sse 0.974481
std 0.017675
2.3 Valid Data:(large sacle)
2.3.1 Traj
2.3.2 XYZ
2.3.3 RPY
2.3.4 APE
1
2
3
4
5
6
7
max 2.925505
mean 0.959726
median 0.750311
min 0.205600
rmse 1.168728
sse 4457.011678
std 0.666971
2.3.5 RPE
1
2
3
4
5
6
7
max 0.372968
mean 0.095626
median 0.073501
min 0.001336
rmse 0.127043
sse 52.648051
std 0.083639
2.4 Valid Data:(7_left_turn_speed_25)
2.4.1 Traj
2.4.2 XYZ
2.4.3 RPY
2.4.4 APE
1
2
3
4
5
6
7
max 0.906424
mean 0.433745
median 0.458177
min 0.032235
rmse 0.482597
sse 44.483811
std 0.211576
2.4.5 RPE
1
2
3
4
5
6
7
max 0.122539
mean 0.059356
median 0.059501
min 0.006692
rmse 0.065355
sse 0.811532
std 0.027350
2.5 Valid Data:(8_right_turn_speed_20)
2.5.1 Traj
2.5.2 XYZ
2.5.3 RPY
2.5.4 APE
1
2
3
4
5
6
7
max 3.209363
mean 0.748066
median 0.543427
min 0.010104
rmse 1.007491
sse 163.421082
std 0.674860
2.5.5 RPE
1
2
3
4
5
6
7
max 1.053835
mean 0.117616
median 0.071761
min 0.011882
rmse 0.194937
sse 6.080060
std 0.155457
2.6 Valid Data:(9_right_turn_speed_40)
2.6.1 Traj
2.6.2 XYZ
2.6.3 RPY
2.6.4 APE
1
2
3
4
5
6
7
max 1.115413
mean 0.527681
median 0.531853
min 0.126005
rmse 0.557352
sse 31.685443
std 0.179429
2.6.5 RPE
1
2
3
4
5
6
7
max 0.443761
mean 0.118309
median 0.098930
min 0.012560
rmse 0.149402
sse 2.254411
std 0.091235
2.7 Valid Data:(10_right_turn_speed_40)
2.7.1 Traj
2.7.2 XYZ
2.7.3 RPY
2.7.4 APE
1
2
3
4
5
6
7
max 1.584476
mean 0.527325
median 0.449033
min 0.076434
rmse 0.602559
sse 52.646241
std 0.291557
2.7.5 RPE
1
2
3
4
5
6
7
max 1.256550
mean 0.229395
median 0.141439
min 0.008515
rmse 0.368821
sse 19.588214
std 0.288803
