BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Asia/Singapore BEGIN:DAYLIGHT DTSTART:20380119T111407 TZOFFSETFROM:+0800 TZOFFSETTO:+0800 RRULE:FREQ=YEARLY;BYDAY=3TU;BYMONTH=1 TZNAME:+08 END:DAYLIGHT BEGIN:STANDARD DTSTART:19820101T000000 TZOFFSETFROM:+0730 TZOFFSETTO:+0800 RRULE:FREQ=YEARLY;BYDAY=1FR;BYMONTH=1 TZNAME:+08 END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20250209T090224Z UID:90460E5F-293B-426F-8615-BFD13B8B4FE9 DTSTART;TZID=Asia/Singapore:20250106T100000 DTEND;TZID=Asia/Singapore:20250106T110000 DESCRIPTION:Title: An LLM-Based Optimization for Integrated Sensing and Com munications in UAV Networks\n\nAbstract: Beyond the fifth-generation (B5G) mobile networks have envisioned future mobile networks to offer not only seamless communication services but also advanced sensing capabilities. Th is has driven interest in joint communication and sensing (C&S) within int egrated sensing and communication (ISAC) systems. Thanks to the high-quali ty line-of-sight (LoS) air-to-ground (A2G) communication links and flexibl e mobility of unmanned aerial vehicles (UAV)\, UAVs are seen as a key enab ler for extending coverage and enhancing C&S performance in ISAC systems. This work focuses on a UAV network with ISAC\, where multiple UAVs simulta neously sense ground user locations with radars and provide communication services. To find the trade-off between C&S\, we formulate a multi-objecti ve optimization problem (MOP) aimed at maximizing both network utility and the localization accuracy (Cramér-Rao bounds\, CRB) of ground users thro ugh optimizing UAV deployment and power control. To solve this highly non- convex problem\, inspired by the huge potential of large language models ( LLM) for prediction and inference\, we introduce an LLM-enabled decomposit ion-based multi-objective evolutionary algorithm (LEDMA). This approach br eaks the MOP into smaller sub-problems and incorporates LLMs as search ope rators using designed prompt engineering within a multi-objective evolutio nary algorithm (MOEA) framework. Numerical results demonstrate that LEDMA effectively identifies the trade-off between C&S and outperforms baseline MOEAs in terms of solution quality and convergence.\n\nBiography: Ming Xia o received received Ph.D degree from Chalmers University of technology\, S weden in November 2007. From November 2007 to now\, he has been in the dep artment of information science and engineering\, school of electrical engi neering and computer science\, Royal Institute of Technology\, Sweden\, wh ere he is currently a full Professor. He was an Editor for IEEE Transactio ns on Communications (2012-2017)\, IEEE Communications Letters (Senior Edi tor Since January 2015) and IEEE Wireless Communications Letters (2012-201 6)\, and has been an Editor for IEEE Transactions on Wireless Communicatio ns since 2018. He has been an area editor for IEEE Open Journal of the Com munication Society since 2019. He received IEEE Vehicular Technology Socie ty Best Magazine Paper Award 2023.\n\nRoom: S2.2-B2-53\, Bldg: NTU\, EEE E xecutive Seminar Rm (S2.2-B2-53) \, Singapore\, Singapore\, Singapore\, 63 9798 LOCATION:Room: S2.2-B2-53\, Bldg: NTU\, EEE Executive Seminar Rm (S2.2-B2-5 3) \, Singapore\, Singapore\, Singapore\, 639798 ORGANIZER:chau.yuen@ntu.edu.sg SEQUENCE:10 SUMMARY:Seminar: An LLM-Based Optimization for Integrated Sensing and Commu nications in UAV Networks URL;VALUE=URI:https://events.vtools.ieee.org/m/456007 X-ALT-DESC:Description: <br /><p class="MsoNormal"><u><span lang="EN-US">Ti tle</span></u><span lang="EN-US">: <strong>An LLM-Based Optimization for I ntegrated Sensing and Communications in UAV Networks</strong></span></p>\n <p class="MsoNormal"><strong><span lang="EN-US">&nbsp\;</span></strong></p >\n<p class="MsoNormal" style="mso-layout-grid-align: none\; text-autospac e: none\;"><u><span lang="EN-US" style="mso-bidi-font-size: 10.5pt\; font- family: 'DengXian Light'\; mso-ascii-theme-font: major-fareast\; mso-farea st-theme-font: major-fareast\; mso-hansi-theme-font: major-fareast\;">Abst ract</span></u><span lang="EN-US" style="mso-bidi-font-size: 10.5pt\; font -family: 'DengXian Light'\; mso-ascii-theme-font: major-fareast\; mso-fare ast-theme-font: major-fareast\; mso-hansi-theme-font: major-fareast\;">: < /span><span lang="EN-US" style="font-family: 'Calibri'\,sans-serif\; color : black\;">Beyond the fifth-generation (B5G) mobile networks<span class="a pple-converted-space">&nbsp\;</span>have envisioned future mobile networks &nbsp\;to offer not only seamless communication services but also advanced sensing capabilities. This has&nbsp\;driven interest in joint communicati on and sensing (C&amp\;S) within integrated sensing and communication (ISA C) systems.&nbsp\;Thanks to the high-quality&nbsp\;line-of-sight (LoS) air -to-ground (A2G) communication links and flexible mobility of&nbsp\;unmann ed aerial vehicles (UAV)\, UAVs&nbsp\;are seen as a key enabler for extend ing coverage and enhancing C&amp\;S performance in ISAC systems.&nbsp\;Thi s work focuses on a UAV network with ISAC\, where multiple UAVs simultaneo usly sense ground user locations with radars&nbsp\;and provide communicati on services. To find the trade-off between C&amp\;S\, we formulate a multi -objective optimization problem (MOP) aimed at maximizing both network uti lity and the localization accuracy (Cram&eacute\;r-Rao bounds\, CRB) of gr ound users through optimizing UAV deployment and power control.&nbsp\;To s olve this highly non-convex&nbsp\;problem\, inspired by the huge potential of large language models (LLM) for prediction and inference\,&nbsp\;we in troduce an LLM-enabled decomposition-based multi-objective evolutionary al gorithm (LEDMA). This approach breaks the MOP into smaller sub-problems an d incorporates LLMs as search operators using designed prompt engineering within a multi-objective evolutionary algorithm (MOEA) framework.&nbsp\;Nu merical results demonstrate that LEDMA effectively identifies the trade-of f between C&amp\;S and outperforms baseline MOEAs in terms of solution qua lity and convergence.</span><span lang="EN-US" style="mso-bidi-font-size: 10.5pt\; font-family: 'DengXian Light'\; mso-ascii-theme-font: major-farea st\; mso-fareast-theme-font: major-fareast\; mso-hansi-theme-font: major-f areast\; mso-bidi-font-family: 'Times New Roman'\; mso-font-kerning: 0pt\; "><span style="mso-spacerun: yes\;">&nbsp\;&nbsp\; </span></span></p>\n<p class="MsoNormal" style="mso-layout-grid-align: none\; text-autospace: non e\;"><span lang="EN-US" style="mso-bidi-font-size: 10.5pt\; font-family: ' DengXian Light'\; mso-ascii-theme-font: major-fareast\; mso-fareast-theme- font: major-fareast\; mso-hansi-theme-font: major-fareast\; mso-bidi-font- family: 'Times New Roman'\; mso-font-kerning: 0pt\;">&nbsp\;</span></p>\n< p class="MsoNormal" style="mso-layout-grid-align: none\; text-autospace: n one\;"><u><span lang="EN-US" style="mso-bidi-font-size: 10.5pt\; font-fami ly: 'DengXian Light'\; mso-ascii-theme-font: major-fareast\; mso-fareast-t heme-font: major-fareast\; mso-hansi-theme-font: major-fareast\; mso-bidi- font-family: 'Times New Roman'\; mso-font-kerning: 0pt\;">Biography</span> </u><span lang="EN-US" style="mso-bidi-font-size: 10.5pt\; font-family: 'D engXian Light'\; mso-ascii-theme-font: major-fareast\; mso-fareast-theme-f ont: major-fareast\; mso-hansi-theme-font: major-fareast\; mso-bidi-font-f amily: 'Times New Roman'\; mso-font-kerning: 0pt\;">: </span><span lang="E N-US">Ming Xiao received received Ph.D degree from Chalmers University of technology\, Sweden in November 2007. From November 2007 to now\, he has b een in the department of information science and engineering\, school of e lectrical engineering and computer science\, Royal Institute of Technology \, Sweden\, where he is currently a full Professor. He was an Editor for I EEE Transactions on Communications (2012-2017)\, IEEE Communications Lette rs (Senior Editor Since January 2015) and IEEE Wireless Communications Let ters (2012-2016)\, and has been an Editor for IEEE Transactions on Wireles s Communications since 2018. He has been an area editor for IEEE Open Jour nal of the Communication Society since 2019. He received IEEE Vehicular Te chnology Society Best Magazine Paper Award 2023.</span></p> END:VEVENT END:VCALENDAR