Master Thesis Open Access

A Fast AI-Based Track Reconstruction on FPGA for the PANDA Experiment

Greta Heine

MARC21 XML Export

<?xml version='1.0' encoding='UTF-8'?>
<record xmlns="http://www.loc.gov/MARC21/slim">
  <leader>00000nam##2200000uu#4500</leader>
  <datafield tag="542" ind1=" " ind2=" ">
    <subfield code="l">open</subfield>
  </datafield>
  <datafield tag="856" ind1="4" ind2=" ">
    <subfield code="s">9020387</subfield>
    <subfield code="u">https://publish.etp.kit.edu/record/22136/files/Master_Thesis_Greta_Heine-final.pdf</subfield>
    <subfield code="z">md5:0ff99df6f2cce1e7933f7ef0b8e1ff9f</subfield>
  </datafield>
  <datafield tag="245" ind1=" " ind2=" ">
    <subfield code="a">A Fast AI-Based Track Reconstruction on FPGA for the PANDA Experiment</subfield>
  </datafield>
  <datafield tag="980" ind1=" " ind2=" ">
    <subfield code="a">thesis</subfield>
    <subfield code="b">master-thesis</subfield>
  </datafield>
  <controlfield tag="001">22136</controlfield>
  <datafield tag="100" ind1=" " ind2=" ">
    <subfield code="a">Greta Heine</subfield>
    <subfield code="u">KIT/ETP, KIT/IPE</subfield>
  </datafield>
  <datafield tag="520" ind1=" " ind2=" ">
    <subfield code="a">&lt;p&gt;Efficient reconstruction of charged particle trajectories is a crucial yet very difficult step&amp;nbsp;in the analysis pipeline of high energy physics (HEP) experiments. Recent work has&amp;nbsp;shown that graph neural networks (GNNs) are well suited for the pattern recognition&amp;nbsp;task of track finding, where tracking detector hits can be naturally represented as nodes&amp;nbsp;and particle track segments as edges. The interaction network (IN) GNN architecture&amp;nbsp;provides computationally efficient edge classification of the high-dimensional and sparse&amp;nbsp;tracker data, which is especially crucial for implementation in constrained computing&amp;nbsp;environments such as field programmable gate arrays (FPGAs). This work describes the&amp;nbsp;overall workflow for implementing and systematically analyzing an IN-based classification&amp;nbsp;of track segments on FPGAs for the anti-Proton&amp;nbsp;Annihilation at&amp;nbsp;DArmstadt (PANDA)&amp;nbsp;forward tracking system (FTS). This workflow includes data preprocessing, graph building, GNN-based edge classification and a series of FPGA implementation design studies&amp;nbsp;concerning latency, resource utilization, and classification quality using the high level&amp;nbsp;synthesis for machine learning (hls4ml) compiler. The presented final implementation of&amp;nbsp;the GNN-based track segment classifier on a Xilinx Zynq&amp;nbsp;provides an overall inference latency of about&amp;nbsp;0.99&amp;nbsp;&amp;mu;s&amp;nbsp;using about&amp;nbsp;34 %&amp;nbsp;of available digital signal processors (DSPs) and&amp;nbsp;85 %&amp;nbsp;of available lookup tables (LUTs). This work enables the acceleration of charged particle tracking on heterogeneous computational resources toward real-time track reconstruction for the PANDA experiment. The discussed methods and studies could be easily adapted and used in other HEP experiments for accelerated charged particle tracking.&lt;/p&gt;</subfield>
  </datafield>
  <datafield tag="260" ind1=" " ind2=" ">
    <subfield code="c">2022-09-15</subfield>
  </datafield>
  <datafield tag="980" ind1=" " ind2=" ">
    <subfield code="a">user-etp</subfield>
  </datafield>
  <datafield tag="980" ind1=" " ind2=" ">
    <subfield code="a">user-hardware</subfield>
  </datafield>
  <controlfield tag="005">20221202161804.0</controlfield>
  <datafield tag="653" ind1=" " ind2=" ">
    <subfield code="a">HEP, Machine Learning, GNN, FPGA, PANDA, Tracking</subfield>
  </datafield>
  <datafield tag="909" ind1="C" ind2="O">
    <subfield code="o">oai:publish.etp.kit.edu:22136</subfield>
    <subfield code="p">user-hardware</subfield>
    <subfield code="p">user-etp</subfield>
  </datafield>
  <datafield tag="700" ind1=" " ind2=" ">
    <subfield code="a">Prof. Dr. Torben Ferber</subfield>
    <subfield code="u">KIT/ETP</subfield>
    <subfield code="4">ths</subfield>
  </datafield>
  <datafield tag="700" ind1=" " ind2=" ">
    <subfield code="a">Dr. Michele Caselle</subfield>
    <subfield code="u">KIT/IPE</subfield>
    <subfield code="4">ths</subfield>
  </datafield>
</record>
Publication date:
September 15, 2022
Keyword(s):
HEP, Machine Learning, GNN, FPGA, PANDA, Tracking
Awarding University:
Karlsruhe Institute of Technology (KIT)
Report Number:
ETP-KA/2022-14
Communities:

Cite as

Export