<?xml version="1.0" encoding="utf-8"?>
<export-example>
  <doc>
    <id>1768</id>
    <completedYear>2020</completedYear>
    <publishedYear/>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>57</pageFirst>
    <pageLast>71</pageLast>
    <pageNumber>15</pageNumber>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject</type>
    <publisherName>Springer International Publishing</publisherName>
    <publisherPlace>Cham</publisherPlace>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>2021-02-21</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Pose Based Trajectory Forecast of Vulnerable Road Users Using Recurrent Neural Networks</title>
    <abstract language="eng">In this work, we use Recurrent Neural Networks (RNNs) in form of Gated Recurrent Unit (GRU) networks to forecast trajectories of vulnerable road users (VRUs), such as pedestrians and cyclists, in road traffic utilizing the past trajectory and 3D poses as input. The 3D poses represent the postures and movements of limbs and torso and contain early indicators for the transition between motion types, e.g. wait, start, move, and stop. VRUs often only become visible from the perspective of an approaching vehicle shortly before dangerous situations occur. Therefore, a network architecture is required which is able to forecast trajectories after short time periods and is able to improve the forecasts in case of longer observations. This motivates us to use GRU networks, which are able to use time series of varying duration as inputs, and to investigate the effects of different observation periods on the forecasting results. Our approach is able to make reasonable forecasts even for short observation periods. The use of poses improves the forecasting accuracy, especially for short observation periods compared to a solely head trajectory based approach. Different motion types benefit to different extent from the use of poses and longer observation periods.</abstract>
    <parentTitle language="eng">ICPR 2021: Pattern Recognition. ICPR International Workshops and Challenges</parentTitle>
    <identifier type="isbn">978-3-030-68763-2</identifier>
    <identifier type="doi">https://doi.org/10.1007/978-3-030-68763-2_5</identifier>
    <enrichment key="copyright">1</enrichment>
    <enrichment key="opus.source">publish</enrichment>
    <licence>Keine Lizenz - es gilt das deutsche Urheberrecht</licence>
    <author>Viktor Kreß</author>
    <author>Stefan Zernetsch</author>
    <author>Konrad Doll</author>
    <author>Bernhard Sick</author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>3D human trajectory forecast</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Vulnerable road users</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>3D body pose</value>
    </subject>
    <subject>
      <language>deu</language>
      <type>swd</type>
      <value>Fahrerassistenzsystem</value>
    </subject>
    <subject>
      <language>deu</language>
      <type>swd</type>
      <value>Fußgänger</value>
    </subject>
    <subject>
      <language>deu</language>
      <type>swd</type>
      <value>Radfahrer</value>
    </subject>
    <collection role="forschungsschwerpunkte" number="">Artifical Intelligence and Data Science</collection>
    <collection role="forschungsschwerpunkte" number="">Intelligent Mobility</collection>
    <collection role="forschungsschwerpunkte" number="">Intelligent Sensors and Signals</collection>
  </doc>
</export-example>
