BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Australia/Adelaide BEGIN:DAYLIGHT DTSTART:20211003T030000 TZOFFSETFROM:+0930 TZOFFSETTO:+1030 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=10 TZNAME:ACDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20220403T020000 TZOFFSETFROM:+1030 TZOFFSETTO:+0930 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=4 TZNAME:ACST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20211007T022150Z UID:6B24A3CB-48AA-46AF-ADB4-3423F8A447B6 DTSTART;TZID=Australia/Adelaide:20211013T193000 DTEND;TZID=Australia/Adelaide:20211013T203000 DESCRIPTION:Dear IEEE members and guests\,\n\nThe IEEE South Australia Comm unications & Signal Processing (C&SP) chapter is pleased to announce a new tech talk series on sensor fusion and tracking.\n\nThe first talk in the series is scheduled on Oct 13\, 2021 07:30 PM. The speaker is Dr. Hoa Van Nguyen from the University of Adelaide and he will be presenting a seminar on multi-object tracking. Please find below the seminar details.\n\nLocat ion: Virtual via the following meeting link\n\nhttps://unisa.zoom.us/j/837 51211051?pwd=bkNHcFRWM1VXT1NpV3lhN3pDbWMrQT09\n\nPassword (if get asked): Adelaide\n\nTime/Date: Wednesday\, 13 October 2021 - 07:30 PM (Adelaide ti me)\n\nTalk Title: Distributed Multi-Object Tracking Under Limited Field o f View Sensors\n\nAbstract: We consider the challenging problem of trackin g multiple objects using a distributed network of sensors. In the practica l setting of nodes with limited field of views (FoVs)\, computing power an d communication resources\, we develop a novel distributed multi-object tr acking algorithm. To accomplish this\, we first formalise the concept of l abel consistency\, determine a sufficient condition to achieve it and deve lop a novel label consensus approach that reduces label inconsistency caus ed by objects' movements from one node's limited FoV to another. Second\, we develop a distributed multi-object fusion algorithm that fuses local mu lti-object state estimates instead of local multi-object densities. This a lgorithm: i) requires significantly less processing time than multi-object density fusion methods\; ii) achieves better tracking accuracy by conside ring Optimal Sub-Pattern Assignment (OSPA) tracking errors over several sc ans rather than a single scan\; iii) is agnostic to local multi-object tra cking techniques\, and only requires each node to provide a set of estimat ed tracks. Thus\, it is not necessary to assume that the nodes maintain mu lti-object densities\, and hence the fusion outcomes do not modify local m ulti-object densities. Numerical experiments demonstrate our proposed solu tion's real-time computational efficiency and accuracy compared to state-o f-the-art solutions in challenging scenarios. We also release source code at https://github.com/AdelaideAuto-IDLab/Distributed-limitedFoV-MOT for ou r fusion method to foster developments in DMOT algorithms.\n\nSpeaker Bio: Hoa Van Nguyen received his Bachelor degree in Electrical Engineering fro m Portland State University\, Oregon\, the USA in June 2012\, and a PhD de gree in computer science from The University of Adelaide\, in July 2020. H e is currently a Post-Doctoral Research Fellow with the School of Computer Science\, The University of Adelaide. His research interests include sign al processing\, robotics\, multi-object tracking and multi-sensor control. \n\nIf you have any problems with accessing the seminar\, please contact n goc.nguyen@mymail.unisa.edu.au.\n\nKind Regards\,\n\nNgoc Hung Nguyen\, Ph D\n\nIEEE SA C&SP Chapter Chair\n\nVirtual: https://events.vtools.ieee.org /m/284591 LOCATION:Virtual: https://events.vtools.ieee.org/m/284591 ORGANIZER:ngoc.nguyen@mymail.unisa.edu.au SEQUENCE:0 SUMMARY:IEEE SA C&SP Tech Talk Series on Sensor Fusion and Tracking - Talk 1: Distributed Multi-Object Tracking Under Limited Field of View Sensors URL;VALUE=URI:https://events.vtools.ieee.org/m/284591 X-ALT-DESC:Description: <br /><p>Dear IEEE members and guests\,</p>\n<p>The IEEE South Australia Communications &amp\; Signal Processing (C&amp\;SP) chapter is pleased to announce a new tech talk series on sensor fusion and tracking.</p>\n<p>The first talk in the series is scheduled on Oct 13\, 2 021 07:30 PM. The speaker is Dr. Hoa Van Nguyen from the University of Ade laide and he will be presenting a seminar on multi-object tracking. Please find below the seminar details.</p>\n<p>Location: Virtual via the followi ng meeting link</p>\n<p><a href="https://unisa.zoom.us/j/83751211051?pwd=b kNHcFRWM1VXT1NpV3lhN3pDbWMrQT09">https://unisa.zoom.us/j/83751211051?pwd=b kNHcFRWM1VXT1NpV3lhN3pDbWMrQT09</a></p>\n<p>Password (if get asked): Adela ide</p>\n<p>Time/Date: Wednesday\, 13 October 2021 - 07:30 PM (Adelaide ti me)</p>\n<p>Talk Title:&nbsp\; Distributed Multi-Object Tracking Under Lim ited Field of View Sensors</p>\n<p>Abstract: We consider the challenging p roblem of tracking multiple objects using a distributed network of sensors . In the practical setting of nodes with limited field of views (FoVs)\, c omputing power and communication resources\, we develop a novel distribute d multi-object tracking algorithm. To accomplish this\, we first formalise the concept of label consistency\, determine a sufficient condition to ac hieve it and develop a novel label consensus approach that reduces label i nconsistency caused by objects' movements from one node's limited FoV to a nother. Second\, we develop a distributed multi-object fusion algorithm th at fuses local multi-object state estimates instead of local multi-object densities. This algorithm: i) requires significantly less processing time than multi-object density fusion methods\; ii) achieves better tracking ac curacy by considering Optimal Sub-Pattern Assignment (OSPA) tracking error s over several scans rather than a single scan\; iii) is agnostic to local multi-object tracking techniques\, and only requires each node to provide a set of estimated tracks. Thus\, it is not necessary to assume that the nodes maintain multi-object densities\, and hence the fusion outcomes do n ot modify local multi-object densities. Numerical experiments demonstrate our proposed solution's real-time computational efficiency and accuracy co mpared to state-of-the-art solutions in challenging scenarios. We also rel ease source code at <a href="https://github.com/AdelaideAuto-IDLab/Distrib uted-limitedFoV-MOT">https://github.com/AdelaideAuto-IDLab/Distributed-lim itedFoV-MOT</a> for our fusion method to foster developments in DMOT algor ithms.</p>\n<p>Speaker Bio: Hoa Van Nguyen received his Bachelor degree in Electrical Engineering from Portland State University\, Oregon\, the USA in June 2012\, and a PhD degree in computer science from The University of Adelaide\, in July 2020. He is currently a Post-Doctoral Research Fellow with the School of Computer Science\, The University of Adelaide. His rese arch interests include signal processing\, robotics\, multi-object trackin g and multi-sensor control.</p>\n<p>If you have any problems with accessin g the seminar\, please contact <a href="mailto:ngoc.nguyen@mymail.unisa.ed u.au">ngoc.nguyen@mymail.unisa.edu.au</a>.</p>\n<p>Kind Regards\,</p>\n<p> Ngoc Hung Nguyen\, PhD</p>\n<p>IEEE SA C&amp\;SP Chapter Chair</p>\n<p>&nb sp\;</p>\n<p>&nbsp\;</p> END:VEVENT END:VCALENDAR