Datasets and Supporting Materials for the MALIN-ANR 2019 Competition (French national research agency)

This "ZENODO deposit" provides a multiple sensor dataset collected by the CyborgLOC team during the intermediate competition of the Challenge MALIN (MAîtrise de la Localisation INdoor), which is a competition for indoor/outdoor real-time positioning. The sensors, including a GNSS receiver Ublox NEO-M8N, a Realsense D435i stereo camera, three Xsens MTi-300 and one PERSY (PEdestrian Reference SYstem), are mounted on different parts of the subject’s body. The PERSY is a foot-mounted positioning device with a tri-axial accelerometer, a tri-axial gyroscope, a tri-axial magnetometer as well as a GNSS receiver Ublox M8T. The two scenarios are designed in a training center of firefighters CFIS (Fire and Rescue Training Center) in Blois, France to simulate the situation of firefighters during interventions. With total distances around 2 km for each scenario, the travelled trajectories passed through challenging environments including indoor, outdoor, urban canyon. The indoor part contains different stair levels, from the underground up to the 6th floor. The travel modes are vehicles and pedestrians. Several classical activities of firefighters are realized such as walking, running, stair-climbing, side-walking, crawling, passing above/below obstacles, carrying a stretcher, ladder climbing, etc. High accurate ground truth of stationary points and enclosing volumes are provided by the organizers of the competition, i.e., the French Ministry of Defense (DGA: Direction Générale de l’Armement). Provided with raw data, they allow the evaluation of the positioning performances.

To facilitate the use of our dataset under Rosbag format, a toolkit of python scripts named MALIN Data Processing Tools is provided on GitHub (https://github.com/4g-group/malin_data_processing_tools). It allows merging Rosbags, converting Rosbag files to CSV files as well as republishing camera’s topics as decompressed data. Details about these processing tools could be found in the Readme file on the Github page.