Figure 1.
Simulation fusion calculation diagram
Figure 2.
Extended Kalman filter fusion results
Figure 3.
Integrating odometer data
Figure 4.
Experimental scenario
TABLE I.
Experimental Result
| Starting point | Target point | Wheel odometry | Error | Fusion odometry | Error |
| (0,0) | (1.3,0) | (1.347,0.035) | (0.047,0.035) | (1.324,0.029) | (0.024, 0.029) |
| (1.3,-3.25) | (1.391,-3.373) | (0.091,0.123) | (1.386,-3.322) | (0.086, 0.072) | |
| (0,-3.25) | (0.152,-3.387) | (0.152,0.137) | (0.085,-3.316) | (0.085, 0.066) | |
| Mean error | (0.097,0.098) | (0.065, 0.056) | |||
Figure 5.
Laser and visual scanning range
TABLE II.
Local map fusion rules
| 2D excitation Optical radar | Depth camera | ||
| Occupy | empty | Uncertain | |
| Occupied | Occupy | Occupy | Occupy |
| empty | Occupy | empty | empty |
| Uncertain | Occupy | empty | Uncertain |
Figure 6.
Robot model
Figure 7.
Simulation experimental environment
Figure 8.
Comparison of simulation experiments
Figure 9.
Real experimental environment
Figure 10.
Installation height of LiDAR
Figure 11.
Robot departure position
Figure 12.
Real environment mapping results
TABLE III.
Positioning results
| Actual location (m) | Actual pose (°) | Gmapping | Vision + Laser | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Estimated position (m) | Estimated pose (°) | Root mean square error of position (cm) | Attitude error (°) | Estimated position (m) | Estimated pose (°) | Root Mean square error of position (cm) | Pose Error (°) | ||
| (6,0) | 30 | (6.085, 0.079) | 33.149 | 11.604 | 3.149 | (6.059, 0.037) | 31.092 | 5.9 | 1.092 |
| (6,-3) | 90 | (6.094,-3.081) | 94.634 | 12.408 | 4.634 | (6.064,-3.051) | 92.043 | 8.184 | 2.043 |
| (6,-6) | 90 | (6.103,-6.089) | 95.005 | 13.612 | 5.005 | (6.072,-6.063) | 92.729 | 9.567 | 2.729 |
| (0,-6) | 180 | (0.135,-6.112) | 186.024 | 17.541 | 6.024 | (0.089,-6.075) | 184.007 | 11.639 | 4.007 |
| (0,0) | 0 | (0.156, 0.127) | 7.678 | 20.116 | 7.678 | (0.091, 0.08) | 4.96 | 12.117 | 4.96 |