mint-lab / TriangulationToolbox

Triangulation Toolbox

Triangulation Toolbox for MATLAB

Triangulation Toolbox is an open-source project to share algorithms, datasets, and benchmarks for landmark-based localization. It is implemented in MATLAB script language and distributed under Simplified BSD License.

• Homepage: http://sites.google.com/site/sunglok/tt

Installation

1. Download the toolbox through its GitHub ZIP URL
2. Unzip the downloaded file, Triangulation-Toolbox-master.zip, on your target directory
3. Execute run_test_aux or run_test_localize to check its working in MATLAB

File Description

Use help command to know each function in detail, for example, help observe_distance.

• 2D Localization Algorithms: localize2d_*.m
• 3D Localization Algorithms: localize3d_*.m
• Observation Functions: observe_distance.m, observe_distance_relative.m, observe_bearing.m, observe_displacement.m, observe_pose.m
• Random Noise Generators: apply_noise_gauss.m
• Accuracy Criteria: error_position.m, error_orientation.m, error_reprojection.m
• Utility Functions
  • Unit-test: test_is_true.m, test_is_near.m
  • Angular Conversions: tran_rad2deg.m, tran_deg2rad.m, tran_rad2rot.m, tran_rot2rad.m, trim_rad.m
• Scripts for Examples: run_example.m
• Scripts for Unit-test: run_test_aux.m, run_test_localize.m
• Scripts for Performance Evaluation
  • Benchmark with Random Landmarks (with Map Uncertainty and Measurement Noise): run_eval_random.m
  • Benchmark with Roh's Angulation Dataset: run_eval_roh.m
  • Benchmark with UTIAS MRCLAM Dataset: run_eval_mrclam.m
  • Position/Orientaion Estimate Distribution Drawing for Each Benchmark: run_draw_distribution.m
• Real Datasets
  • Roh's Angulation Dataset: dataset_roh
  • UTIAS MRCLAM Dataset: dataset_mrclam (converted by run_conv_mrclam.m)

Example

trueMap =                                      ...
[                                              ...
    % x,  y,  z, r_x, r_y, r_z                 ...
      0,  0,  0,   0,  0,  0;                  ...
      8,  0,  0,   0,  0,  tran_deg2rad( +90); ...
      8,  8,  0,   0,  0,  tran_deg2rad(-180); ...
];
truePose = [3, 2, 0, 0, 0, pi / 9];
obsData = observe_distance(trueMap, truePose);      % Simulate observation
estPose = localize2d_sayed05_toa(obsData, trueMap); % Estimate position

See run_example.m for more complex example with visualization.

Reference

• Sunglok Choi, Triangulation Toolbox: Open-source Algorithms and Benchmark for Landmark-based Localization, in Proceedings of IEEE International Conference on Robotics and Automation (ICRA), 2014

Authors

• Sunglok Choi (sunglok AT hanmail DOT net)

Acknowledgement

The authors thank to the following contributors and projects.

• Prof. Federico Thomas shared his trilateration code at his homepage. (See the algorithm)
• Hyun-Chul Roh provided his angulation dataset. (See the dataset)
• Dr. Keith Leung, Yoni Halpern, Prof. Tim Barfoot, and Prof. Hugh Liu shared their MRCLAM dataset. (See the dataset)