Co-Investigators:
Annotation:
Surveillance systems are an essential part of the critical infrastructure for ensuring the safety and the security of aviation by detecting and tracking targets. The data association in the scope of target tracking solves the issue of assigning measurements to targets. If a radar or other sensor outputs a measurement, it is necessary to determine if it is a new target, a false alarm or an already existing one and if so, to which target does the measurement belong. In practice there are numerous types of sensors whose outputs are anonymous measurements in the form of position or other relevant quantities and where the true identification is not known. In recent years machine learning has received considerable attention especially due to the reduced cost of sufficient computing power. The application of machine learning to the data association in the scope of target tracking has not been thoroughly explored yet. The use of machine learning to improve data association methods, reduce combinatorial explosion and implementation complexity is a path that could provide simplification of this ubiquitous task. Improvement of the data association presents not only an issue with significant research potential but also with wide practical and industrial applicability.
Department:
Year:
2022 - 2023
Program:
Studentská grantová soutěž ČVUT - SGS22/128/OHK2/2T/16