BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//TU/e ME//Events//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:Europe/Amsterdam
BEGIN:STANDARD
DTSTART:19701025T030000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:19700329T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:event-18@tuemeche.nl
DTSTAMP:20261008T003056Z
DTSTART;TZID=Europe/Amsterdam:20260929T133000
DTEND;TZID=Europe/Amsterdam:20260929T150000
SUMMARY:Semantic World Models for Multi-robot systems
DESCRIPTION:Speaker: Koen de Vos\nHost: Elena Torta\n\nMulti-robot system
 s are increasingly deployed in semi-structured environments\nsuch as ware
 houses and dairy farms\, where robots must share space\, and\ncoordinate 
 their motion in a robust\, safe and scalable way. Model Predictive\nContr
 ol (MPC) provides a flexible framework for generating safe trajectories\,
 \nbut it scales poorly with the number of agents\, struggles to represent
  discrete\ncoordination decisions such as yielding\, and depends on relia
 ble and low latency\ninter-agent communication. This thesis proposes that
  the decision-making\nof mobile-robot teams in semi-structured environmen
 ts can be made more\nrobust and computationally tractable by introducing 
 a shared world model\nthat encodes both the spatial structure and geometr
 y of the environment.\nBy automatically configuring the constraints\, obj
 ectives\, and agent groupings\nof individual MPC controllers from this sh
 ared representation\, the need for\ncoupled global planning is reduced wh
 ile safe and deadlock-free operation\nis maintained. Four contributions a
 re presented in this thesis. Chapter 2\nintroduces a property-graph world
  model built from three map primitives\, Areas\,\nBoundaries and Interfac
 es\, and a systematic mapping from these primitives\nto MPC constraints a
 nd objectives\, allowing controllers to be reconfigured\nat runtime and a
 gents to be dynamically grouped only when their semantic\narea horizons o
 verlap. Chapter 3 adds a discrete claim policy on top of the\ncontinuous 
 MPC layer that assigns access rights to areas\, ensures deadlock-free\nop
 eration\, and enables task-dependent area occupancy constraints\; the app
 roach\nis validated in simulation and on physical robot hardware. Chapter
  4 extends\nthe framework to a distributed ADMM-MPC setting in which the 
 world model\nselectively activates inter-agent coupling and yielding deci
 sions. Chapter 5\ncloses the loop by deriving the required world models f
 rom Building Information\nModels using an intermediate representation and
  a sampling-based roadmap to\nrecover navigable zones. Together\, these c
 hapters show that a shared semantic\nworld model is able to configure MPC
  controllers at runtime\, to decouple\nagent problems through discrete co
 ordination\, and to reduce computational\nand communication requirements 
 compared to baseline methods\, leading to\nmulti-robot coordination that 
 is more robust in execution\, and configurable at\nruntime\n\nMore info: 
 https://tuenl.sharepoint.com/sites/intranet-mechanical-engineering/_layou
 ts/15/Event.aspx?ListGuid=9bfaaae6-070c-4371-810d-a43d7ee02bf2&ItemId=243
LOCATION:Atlas 0.710
URL:https://tuemeche.nl/peoplepages/event.php?id=18
CATEGORIES:PhD Defense
END:VEVENT
END:VCALENDAR