Research Data

This data package contains open research data intended for machine learning from the WinOSens project. It is sensor data that was recorded by a sensor module attached to a sack truck pushed on the same path many times. It was meant to train ML models for recognizing the current position on the path. Sensor values are available for acceleration, gyration, magnetic flux density, air pressure, temperature, humidity, and the brightness of different colors. The current location of the sack truck is available as labels annotated by humans. The data is provided for 3 paths that cover indoor and outdoor areas.

The refined dataset consists of edited data from raw AIS data. The data is organised as follows: There is one csv file per cargo ship (271 files in total from 1.csv to 271.csv). This file contains all calls of the respective ship at the Port of Miami for the given period from 2018 to 2020. To anonymize the data, the identifying ship data, i.e. the MMSI, was removed. Further information can be found in the paper in Section 3.1. (paper is under review and will be linked here as soon as available) The raw AIS data which are used are available on the websites of the National Oceanic and Atmospheric Administration (NOAA) Office for Coastal Management. An overview is published here: https://marinecadastre.gov/ais/

This dataset contains synthetic structures of hierarchical nanoporous gold, which mimics real hierarchical nanoporous gold structures, segmentation results and repositioned interpolated structures of the real hierarchical nanoporous gold dataset and trained weights of machine learning model used for interpolation. For more information, please refer to the published research article: Role of slice thickness quantification in the 3D reconstruction of FIB tomography data of nanoporous materials.