BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Inverse Methods for State and Physical Property Estimation of Non-
 Cooperative Space Objects
DTSTART:20261126T133000
DTEND:20261126T153000
DTSTAMP:20261109T193630Z
UID:a2ab682c0ea7c4e742c2453e45d2321132877a1ccd77a507681ab247
CATEGORIES:Conferences - Seminars
DESCRIPTION:Pierre Ancey\nEDIC candidacy exam\nExam president: Prof. Pasca
 l Frossard\nThesis advisor: Prof. Mathieu Salzmann\nThesis co-advisor: Dr.
  Andrew Lawrence Price\nCo-examiner: Prof. Jean-Paul Kneib\n\nAbstract\nTh
 e rapid proliferation of satellites and orbital debris is making Earth's o
 rbital environment increasingly congested\, raising the risk of collisions
  and threatening the long-term sustainability of space operations. Mitigat
 ing these risks requires accurate knowledge of the objects in orbit\, incl
 uding not only their trajectories but also their rotational dynamics\, geo
 metry\, and physical properties. While direct inspection in space could pr
 ovide some of this information\, inspection missions are costly and diffic
 ult to scale to the growing orbital population. Ground-based optical obser
 vations\, such as telescope images and light curves\, offer a comparativel
 y accessible and cost-effective means of monitoring these objects. However
 \, many satellites and debris objects are non-cooperative\, providing no t
 elemetry or prior knowledge of their physical characteristics\, making the
  recovery of such information from optical measurements particularly chall
 enging.\n\nThese observations result from complex interactions between the
  object's physical properties and motion\, illumination conditions\, atmos
 pheric effects\, and the observing instrument. Accurately modeling how the
 se factors produce the measured signal\, known as the forward problem\, is
  already challenging. Recovering the unknown state and physical properties
  from the observations\, the inverse problem\, is considerably more diffic
 ult. Measurements are often sparse\, noisy\, and incomplete\, while differ
 ent combinations of parameters can produce nearly identical observations\,
  making the problem ambiguous and potentially ill-posed.\n\nThis research 
 aims to develop efficient and robust algorithms for estimating the states 
 and physical properties of non-cooperative space objects from optical obse
 rvations. By combining physics-based models with numerical optimization\, 
 probabilistic inference\, and learning-based methods\, the objective is to
  recover information that is difficult to obtain using conventional approa
 ches. Particular attention will be given to exploiting physical constraint
 s\, quantifying uncertainty\, improving computational efficiency\, and bri
 dging the gap between simulated and real observations.\n\nUltimately\, thi
 s research seeks to enable a more complete understanding of spacecraft and
  orbital debris from limited observations\, supporting safer and more sust
 ainable operations in Earth's orbit.\n\nBackground papers\n \n\n	Burton\,
  A.\, Robinson\, L.\, & Frueh\, C. (2024). Light curve attitude estimation
  using particle swarm optimizers. Advances in Space Research\, 74(11)\, 56
 19–5638.\n\n     https://www.sciencedirect.com/science/article
 /pii/S0273117724009281\n \n\n	Radev\, S. T.\, Mertens\, U. K.\, Voss\, A.
 \, Ardizzone\, L.\, & Köthe\, U. (2022). BayesFlow: Learning complex stoc
 hastic models with invertible neural networks. IEEE Transactions on Neural
  Networks and Learning Systems\, 33(4)\, 1452–1466.\n\n       
 https://arxiv.org/abs/2003.06281\n \n\n	Liu\, Y.\, Sun\, Z.\, Zhang\, L.\
 , Xi\, L.\, Dong\, W.\, & Deng\, F. (2025). Spacecraft-NeRF: High-Fidelity
  Reconstruction of Spacecraft by Neural Radiance Field Based Implicit Repr
 esentation. IEEE Transactions on Aerospace and Electronic Systems\, 61(6)\
 , 15182–15194.\n\n       https://ieeexplore.ieee.org/document/
 11260637\n 
LOCATION:BC 233 https://plan.epfl.ch/?room==BC%20233
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR