ipin competition 2016 track 4
play

IPIN competition 2016 - Track 4 Sponsored by Indoor Mobile Robot - PowerPoint PPT Presentation

IPIN competition 2016 - Track 4 Sponsored by Indoor Mobile Robot Positioning http://asti.es/en/ Chairs: M Carmen Prez Cristina Losada Co-Chairs: Felipe Espinosa Javier Macias-Guarasa International Conference on Indoor Positioning and


  1. IPIN competition 2016 - Track 4 Sponsored by Indoor Mobile Robot Positioning http://asti.es/en/ Chairs: Mª Carmen Pérez Cristina Losada Co-Chairs: Felipe Espinosa Javier Macias-Guarasa International Conference on Indoor Positioning and Indoor Navigation ] organised by EvAAL

  2. TRACK 4 PRES ENTATION  GOAL: est imate the traj ectory followed by a mobile robot along a predetermined track inside an indoor area.  COMPETITION S ET-UP:  Robot: EAS YBOT of AS TI  Magnetic guidance system.  The robot speed does not exceed 0.5m/ s, but it is not uniform during the traj ectory.  Reference point t o be localized 0.34m in the middle of the robot. 1.64m  No possibility of interaction with the mobile robot. 0.52m  The traj ectory (ground truth) is not able to be seen for competitors.

  3. TRACK 4 PRES ENTATION  COMPETITION S ET-UP:  Navigation area:  Area 12m x 6m.  4 strut profiles where participants can place their equipment. Adj ustable height up to 315cm.  No limitation in the number of sensors to be located in the strut profiles.  3 positions provided in advance for calibrating purposes, including the (0,0) of the robot starting point.  30 minutes set-up (45min in practice). Strut profiles  Test duration: 3-4 minutes. Two loops.  Order of participation by drawing lots.

  4. TRACK 4 PRES ENTATION  Ground truth: 6 5 4 3 Participants have to 2 report the traj ectory: X coordinate (m)  X coordinate (mm) 1  Y coordinate (mm) 0  Timestamp (ms) -1  Minimum frame-rate: 0.1s -2 -3 -4 -5 -6 4 3 2 1 0 Y coordinate (m)

  5. TRACK 4 PRES ENTATION Robot path Estimated path Euclidean distance  Evaluation method: 4 Y coordinate (m) 3 2 1 0 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 X coordinate (m) 100 90 80 172 mm  Euclidian distance between the estimated 70 positions and the ground truth. % of measurements 60 75%  The order of the measurements is 50 considered. 40 30  Final scores are computed considering the third quartile (CDF at 75% ). 20 10 0 0 200 400 600 800 1000 Error (mm)

  6. Competitors of Track 4  Janis Tiemann, Fabian Eckermann and Christian Wietfeld, “ ATLAS - An Open-Source TDOA-based Ultra-Wideband Localization System ” , TU Dortmund University, Communication Networks Institute (CNI), Germany.  Maxim Katkov and S ergey Huba, “ TOPOrtls TPM UWB wireless system”, R & D Dept, TopoR TLS , Belarus/ S andwich, MA, US A.

  7. Competitors of Track 4- Out of Competition  J. Ureña, J. M. Villadangos, D. Gualda, M. C. Pérez, A. Hernández, J. J. García, A. Jiménez, J. C. García, J. F . Arango, E. Díaz, “ Locate-US: an Ultrasonic Local Positioning System based on Encoded Beacons ” , Electronics Department, University of Alcalá. 100 90 80 4 70 % of measurements 60 Y coordinate (m) 3 50 2 40 30 1 20 0 10 0 0 200 400 600 800 1000 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 Error (mm) X coordinate (m)

  8. Competitors of Track 4- Out of Competition  R. Nieto, E. S antiso, P . López, M. Martínez, F . Espinosa. “ Laser scanner network for indoor positioning and tracking ” , Electronics Department, University of Alcalá. 100 90 80 70 % of measurements 4 60 Y coordinate (m) 3 50 40 2 30 1 20 0 10 0 0 200 400 600 800 1000 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 Error (mm) X coordinate (m)

  9. RANKING 4 Y coordinate (m) 3 2 1 0 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 X coordinate (m) 4 Y coordinate (m) 3 2 1 0 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 X coordinate (m)

  10. RANKING  Cumulative Distribution Function of the Euclidian error 1 0.9 X1 X2 0.8 106 mm 311 mm 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 100 200 300 400 500 600 700 800 900 1000 Error (mm) Participant CDF 75% Average Error X1 106mm 80.12mm X2 311mm 220.46mm

  11. THE WINNER 1500€ cash prize CDF(75%)=106mm Janis Tiemann, Fabian Eckermann, and Christian Wietfled

Recommend


More recommend