MagPIE Dataset

MagPIE is a publicly available dataset for the evaluation of indoor positioning algorithms that use magnetic anomalies. Our dataset contains IMU and magnetometer measurements along with ground truth position measurements that have centimeter-level accuracy. To produce this dataset, we collected over 13 hours of data (51 kilometers of total distance traveled) from three different buildings, with sensors both handheld and mounted on a wheeled robot, in environments with and without changes in the placement of objects that affect magnetometer measurements ("live loads'').

Dataset and Code

The file structure of the MagPIE dataset.

Full dataset with all 723 sequences: zip (545MB)

Data for computing IMU Allan variance: zip (5.36MB)

ROS package to convert individual MagPIE trials to ROS bags: click here

Dataset Characteristics

  
    	Building and Platform
    	Num. Training Cases
    	Num. Dead Load Cases
    	Num. Live Load Cases
  

  	    CSL UGV
        93
    	11
    	10
  

  	    CSL WLK
        100
    	11
    	11
  

  	    Talbot UGV
        100
    	11
    	11
  

  	    Talbot WLK
        99
    	11
    	11
  

  	    Loomis UGV
        99
    	11
    	11
  

  	    Loomis WLK
        100
    	12
    	11
  

  
    	Aggregating over all trials:
  
  	    Total Time
        13.96 hrs
  
  	    Total Distance
        51.35 km
  
  	    Average Time of Test Cases
        90.80 sec
  
  	    Minimum Time of Test Cases
        21.10 sec
  
  	    Maximum Time of Test Cases
        245.66 sec
  
  	    Average Distance of Test Cases
        101.54 m
  
  	    Minimum Distance of Test Cases
        23.01 m
  
  	    Maximum Distance of Test Cases
        297.98 m

We created an Android application for the Lenovo Phab 2 Pro which uses Google Tango.

We also used an Android application to store and file IMU and magnetometer data on a Moto Z Play.

Publications

D. Hanley, A. B. Faustino, S. D. Zelman, D. A. Degenhardt, and T. Bretl, "MagPIE: A Dataset for Positioning with Magnetic Anomalies", in Eigth International Conference of Indoor Positioning and Indoor Navigation (IPIN), Sapporo, Japan, September 2017. [pdf]

Support

Please send bug reports to David Hanley: hanley6@illinois.edu

KNown issues

7/21/2017 - The relative orientations of the Z Play and Phab 2 Pro for the UGV and WLK cases are different causing issues for mapping based on individual axes measurements. FIXED: README now includes correct rotations matrix for UGV and WLK cases.