BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:ESE defense | Hongchao Zhang
X-WR-TIMEZONE:Central Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260721T173502Z
UID:tag:localist.com\,2008:EventInstance_49783154731986
DTSTART:20250617T150000Z
DTEND:20250617T170000Z
DESCRIPTION:Title: Resilient Safe Control of Autonomous Systems\n\nAbstract
 : Asimov's Three Laws of Robotics famously outlined fundamental safety pri
 nciples governing robot-human interaction. This foundational concept of sa
 fety is paramount for today's autonomous systems\, such as robots\, which 
 possess inherent cyber-physical properties. With the increasingly widespre
 ad application of Autonomous systems in real-world environments\, the chal
 lenges facing research on formal safety verification have grown even more 
 significant. However\, end-to-end verification of such complex\, integrate
 d systems remains an open and formidable challenge due to their high dimen
 sionality\, nonlinearity\, and the use of learning-based components. This 
 thesis approaches this challenge by pursuing verifiably safe autonomy from
  two complementary directions: (i) safe control of learning-enabled system
 s providing formal guarantees and (ii) resilient safe control that maintai
 ns formal safety guarantees under extreme scenarios such as sensor faults 
 and cyber-physical attacks.\n\nThe first half of this dissertation present
 s the formal verification of autonomous systems that integrate learning-en
 abled components. It starts with the safety verification of neural control
  barrier functions (NCBF) employing Rectified Linear Unit (ReLU) activatio
 n functions. By leveraging a generalization of Nagumo's theorem\, we propo
 se exact safety conditions for deterministic systems. To manage computatio
 nal complexity\, we enhance the efficiency of verification and synthesis u
 sing a VNN-based (Verification of Neural Networks) search algorithm and a 
 neural breadth-first search algorithm. We further propose the synthesis an
 d verification of safe control for stochastic systems.\n\nThe second half 
 of this dissertation broadens the scope of end-to-end verification by expl
 icitly accounting for imperfections and perturbations. We first proposed F
 ault-Tolerant Stochastic CBFs and NCBFs to provide safety guarantees for a
 utonomous systems under state estimation error caused by low-dimensional s
 ensor faults and attacks. We then investigate the unique challenges posed 
 by Light Detection And Ranging (LiDAR) perception attacks. We propose a fa
 ult detection\, identification\, and isolation mechanism for 2D and 3D LiD
 AR and provide safe control under attacks.
LOCATION:McKelvey Hall\, 1030
SUMMARY:ESE defense | Hongchao Zhang
URL;VALUE=URI:https://happenings.washu.edu/event/ese-defense-hongchao-zhang
CATEGORIES:Seminar/Colloquia
END:VEVENT
END:VCALENDAR