VRU Trajectory Dataset

<p><strong>VRUT_Dataset_complete.tar.gz</strong> :</p> <p>The VRU Trajectory Dataset consists of 1068 pedestrian and 464 cyclist trajectories recorded at an urban intersection using cameras and LiDARs. A detailed description of the intersection can be found in [1]. The pedestrian trajectories were recorded by using a wide angle stereo camera system to track the pedestrians&#39; head position and generating the 3D position by triangulation. The cyclists trajectories were recorded by using LiDARs to track the center of gravity of the cyclists. The cameras operate at 50 Hz, the LiDARs at 12.5 Hz. The dataset partly results from the projects DeCoInt&sup2; [2] funded by the &quot;German Reasearch Foundation&quot; (DFG) and AFUSS funded by the &quot;Bundesministerium f&uuml;r Bildung und Forschung&quot; (BMBF). Additionally, our work is supported by &quot;Zentrum Digitalisierung Bayern&quot;.</p> <p>Dataset Format:</p> <p>The complete dataset consists of 1532 files in csv format, where every file contains one VRU trajectory. A csv file consists of 4 columns:</p> <ul> <li>ID: Measurement IDs</li> <li>timestamp: Timestamp in seconds</li> <li>x: position in x-direction in meter</li> <li>y: position in y-direction in meter<br> &nbsp;</li> </ul> <p><strong>extended_cyclist_dataset.7z</strong> :</p> <p>To create the Extended Cyclist Trajectory Dataset, we changed the camera lenses to capture a wider filed of view enclosing the bike lane. The dataset currently consists of 1 746 cyclist trajectories including motion primitive labels. The motion primitve labels include the classes <em>wait</em>, which starts at the last visbible movement of the bicycle wheel and ends the first visible bicycle wheel, <em>starting movement, </em>which is the frame of the first visible movement of the cyclist before the end of <em>wait, </em><em>tr/tl </em>(turn left, turn right), which start at the first and end at the last visible turning movement of the cyclist, <em>and hand signal left/right, </em>which start at the first and end at the last visible frame of the hand signal.</p> <p>Dataset Format</p> <p>The dataset consists of 1746 files in json format, where every file contains one cyclist trajectory.</p> <p>A json file is structured as follows:</p> <p>{&quot;vru_type&quot;: &quot;bike&quot;,<br> &nbsp; &quot;trajectory2: [{&quot;Timestamp&quot;: [LIST OF UTC TIMESTAMPS],<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &quot;x&quot;: [LIST OF X POSISTIONS],<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &quot;y&quot;: [LIST OF Y POSISTIONS],<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &quot;z&quot;: [LIST OF Z POSISTIONS],<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &quot;x_smoothed&quot;: [LIST OF SMOOTHED X POSISTIONS],<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &quot;y_smoothed&quot;: [LIST OF SMOOTHED Y POSISTIONS],<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &quot;z_smoothed&quot;: [LIST OF SMOOTHED Z POSISTIONS]}],<br> &nbsp; &quot;motion_primitives&quot;: {&quot;mp_labels&quot;:<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; [{&quot;mp_label&quot;: LABEL NAME,<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &quot;start_time&quot;: START UTC TIMESTAMP,<br> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &quot;end_time&quot;: END UTC TIMESTAMP, ...}, ...]}}</p> <p><strong>cyclist_starting_dataset.zip</strong>:</p> <p>The complete dataset is stored in csv format. It contains the sensor data and the corresponding label. The sensor data csv file consists of 19 columns:</p> <p>The data is split into meta data and payload data. The meta data contains the fields identifying<br> an experiment, i.e., ExperimentID, SceneID, VRUID. Additionally, the Timestamp field is used to identify the corresponding sensor readings.</p> <ul> <li>ExperimentID: meta data, identifies each different experiment</li> <li>SceneID: meta data</li> <li>VRUID: meta data, identifier for test subject</li> <li>Timestamp: meta data, Timestamp, starting at 0 for each separate experiment</li> </ul> <p>An experiment always contains one VRU but it may consist of several scenes, i.e., starting movements. In the following the fields containing the sensor values are described.</p> <ul> <li>Accelerometer_x: Acceleration force along the x axis (including gravity), in m/s&sup2;</li> <li>Accelerometer_y: Acceleration force along the y axis (including gravity), in m/s&sup2;</li> <li>Accelerometer_z: Acceleration force along the z axis (including gravity), in m/s&sup2;</li> <li>Linear_Accelerometer_x: Acceleration force along the x axis (excluding gravity), in m/s&sup2;</li> <li>Linear_Accelerometer_y: Acceleration force along the y axis (excluding gravity), in m/s&sup2;</li> <li>Linear_Accelerometer_z: Acceleration force along the z axis (excluding gravity), in m/s&sup2;</li> <li>Gyroscope_x: Rate of rotation around the x axis, in rad/s</li> <li>Gyroscope_y: Rate of rotation around the y axis, in rad/s</li> <li>Gyroscope_z: Rate of rotation around the z axis, in rad/s</li> <li>Rotation_w: Scalar component of the rotation vector</li> <li>Rotation_x: Rotation vector component along the x axis, unitless</li> <li>Rotation_y: Rotation vector component along the y axis, unitless</li> <li>Rotation_z: Rotation vector component along the z axis, unitless</li> <li>target: indicates the label. Nothing = no label, 0 = waiting, 1 = starting_movement, 2 = starting</li> </ul> <p>&nbsp;</p> <p>This work results from the project DeCoInt 2, supported by the German Research Foundation (DFG) within the priority program SPP 1835: &quot;Kooperativ interagierende Automobile&quot;, grant numbers DO 1186/1-2, FU 1005/1-2, and SI 674/11-2. Additionally, the work is supported by &quot;Zentrum Digitalisierung Bayern&quot;.</p> <p><strong>References</strong></p> <table> <tbody> <tr> <td>[1] M. Goldhammer, E. Strigel, D. Meissner, U. Brunsmann, K. Doll and K. Dietmayer, &quot;Cooperative multi sensor network for traffic safety applications at intersections,&quot; 2012 15th International IEEE Conference on Intelligent Transportation Systems, Anchorage, AK, 2012, pp. 1178-1183. doi: 10.1109/ITSC.2012.6338672</td> </tr> <tr> <td>[2] M. Bieshaar, G. Reitberger, S. Zernetsch, B. Sick, E. Fuchs and K. Doll, &quot;Detecting Intentions of Vulnerable Road Users Based on Collective Intelligence&quot;, AAET &ndash; Automatisiertes und vernetztes Fahren, Braunschweig, pp. 67-87, Available: https://www.its-mobility.de/download/AAET/Dokumentation/AAET_2017_Tagungsband_Download.pdf</td> </tr> <tr> <td>[3] M. Bieshaar, S. Zernetsch, A. Hubert, B. Sick, and K. Doll. Cooperative starting movement detection of cyclists using convolutional neural networks and a boosted stacking ensemble. CoRR, abs/1803.03487, 2018.</td> </tr> </tbody> </table>