• Acta Optica Sinica
  • Vol. 39, Issue 5, 0528003 (2019)
Rendong Wang*, Hua Li, Kai Zhao, and Youchun Xu**
Author Affiliations
  • Army Military Transportation University, Tianjin 300161, China
  • show less
    DOI: 10.3788/AOS201939.0528003 Cite this Article Set citation alerts
    Rendong Wang, Hua Li, Kai Zhao, Youchun Xu. Robust Localization Based on Kernel Density Estimation in Dynamic Diverse City Scenes Using Lidar[J]. Acta Optica Sinica, 2019, 39(5): 0528003 Copy Citation Text show less
    High-accuracy maps constructed by proposed algorithm. (a) Height map; (b) reflectivity map
    Fig. 1. High-accuracy maps constructed by proposed algorithm. (a) Height map; (b) reflectivity map
    Flow chart of proposed localization algorithm
    Fig. 2. Flow chart of proposed localization algorithm
    Satellite map of experimental section
    Fig. 3. Satellite map of experimental section
    Moving trajectories before and after localization by proposed algorithm
    Fig. 4. Moving trajectories before and after localization by proposed algorithm
    Comparison of localization results in experiment 1 by three algorithms. (a) Lateral error; (b) longitudinal error; (c) heading angle error
    Fig. 5. Comparison of localization results in experiment 1 by three algorithms. (a) Lateral error; (b) longitudinal error; (c) heading angle error
    Localization results of cross road. (a) Aerial view; (b) 3D view; (c) forward camera view; (d) HF algorithm; (e) ML-RANSAC algorithm; (f) proposed algorithm
    Fig. 6. Localization results of cross road. (a) Aerial view; (b) 3D view; (c) forward camera view; (d) HF algorithm; (e) ML-RANSAC algorithm; (f) proposed algorithm
    Localization results of straight road. (a) Aerial view; (b) 3D view; (c) forward camera view; (d) HF algorithm; (e) ML-RANSAC algorithm; (f) proposed algorithm
    Fig. 7. Localization results of straight road. (a) Aerial view; (b) 3D view; (c) forward camera view; (d) HF algorithm; (e) ML-RANSAC algorithm; (f) proposed algorithm
    Horizontal position error distributions by three algorithms under different initial pose deviations. (a) Scatter diagram; (b) histogram
    Fig. 8. Horizontal position error distributions by three algorithms under different initial pose deviations. (a) Scatter diagram; (b) histogram
    Deviation /mAlgorithmHorizontal root-mean-square error /mLongitudinal root-mean-square error /mLateral root-mean-square error /mHeading angleroot-mean-square error /(°)Phoriz /%Time-cost /s
    Lateral error <0.1 mLateral error <0.4 mMeanStandard deviation
    HF0.76480.44130.56510.564736.2385.420.92480.8710
    2.5ML-RANSAC0.21890.15330.12050.219119.1792.930.46340.2934
    Proposed0.21770.19910.04760.141057.0095.310.34490.2394
    HF7.01424.50704.33662.076510.6451.774.912911.2041
    5.0ML-RANSAC1.99421.91820.31660.342720.6387.800.50830.3859
    Proposed0.14110.11980.05100.153954.9195.950.34580.1978
    HF15.8015.462214.41405.563209.0931.31259.6273
    10.0ML-RANSAC2.28802.21570.30600.378117.7680.440.63320.5340
    Proposed0.23200.20010.07490.216753.0593.180.59940.9812
    Table 1. Localization statistical results by three algorithms under different initial pose deviations
    Rendong Wang, Hua Li, Kai Zhao, Youchun Xu. Robust Localization Based on Kernel Density Estimation in Dynamic Diverse City Scenes Using Lidar[J]. Acta Optica Sinica, 2019, 39(5): 0528003
    Download Citation