BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Australia/Canberra BEGIN:DAYLIGHT DTSTART:20261004T030000 TZOFFSETFROM:+1000 TZOFFSETTO:+1100 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=10 TZNAME:AEDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20260405T020000 TZOFFSETFROM:+1100 TZOFFSETTO:+1000 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=4 TZNAME:AEST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260601T230928Z UID:AF10817A-2973-42A3-B5F1-44708110B05E DTSTART;TZID=Australia/Canberra:20260618T140000 DTEND;TZID=Australia/Canberra:20260618T143000 DESCRIPTION:Autonomous robots are increasingly deployed across defence\, in dustrial\, and public-sector settings\, yet remain exposed to cyber-physic al attacks that conventional cybersecurity handles poorly. Standard method s treat a robot as a generic computing endpoint\, ignoring the physical co nstraints that set it apart. The consequence is that a network breach quic kly becomes physical harm through compromised navigation\, manipulated sen sor streams\, or hijacked actuators. This talk presents a physics-informe d approach to securing robotic vehicles\, built on a simple premise: robot s obey physics\, and their kinematic and electromechanical constraints lea ve detectable signatures usable for intrusion detection. The first part de velops detectors that operate within these constraints across several plat forms\, including deep-learning detection of denial-of-service attacks on ROS 2 ground vehicles\, power-consumption side channels for man-in-the-mid dle attacks on differential-drive robots\, and bio-mechanical coupling bet ween wing oscillation and inertial measurements on flapping-wing micro air vehicles. These methods achieve high detection performance while remainin g light enough for microcontroller and single-board deployment. The second part outlines the direction for the remaining programme: assessing how re silient physics-informed detectors are to adversarial evasion crafted with in physical constraints\, and developing defences that harden them against adaptive adversaries. The talk closes with the open questions and benchma rks needed to move from detection toward trustworthy\, attack-resilient au tonomy.\n\nCo-sponsored by: Dr Fendy Santoso\n\nVirtual: https://events.vt ools.ieee.org/m/561672 LOCATION:Virtual: https://events.vtools.ieee.org/m/561672 ORGANIZER:m.garratt@unsw.edu.au SEQUENCE:32 SUMMARY:Trusted Operations of Robotic Systems in Contested Environments URL;VALUE=URI:https://events.vtools.ieee.org/m/561672 X-ALT-DESC:Description: <br /><p class="MsoNormal"><span style="mso-fareast -font-family: 'Times New Roman'\; color: black\;">Autonomous robots are in creasingly deployed across defence\, industrial\, and public-sector settin gs\, yet remain exposed to cyber-physical attacks that conventional cybers ecurity handles poorly. Standard methods treat a robot as a generic comput ing endpoint\, ignoring the physical constraints that set it apart. The co nsequence is that a network breach quickly becomes physical harm through c ompromised navigation\, manipulated sensor streams\, or hijacked actuators . </span><span style="mso-fareast-font-family: 'Times New Roman'\; color: black\;">&nbsp\;</span><span style="mso-fareast-font-family: 'Times New Ro man'\; color: black\;">This talk presents a physics-informed approach to s ecuring robotic vehicles\, built on a simple premise: robots obey physics\ , and their kinematic and electromechanical constraints leave detectable s ignatures usable for intrusion detection. The first part develops detector s that operate within these constraints across several platforms\, includi ng deep-learning detection of denial-of-service attacks on ROS 2 ground ve hicles\, power-consumption side channels for man-in-the-middle attacks on differential-drive robots\, and bio-mechanical coupling between wing oscil lation and inertial measurements on flapping-wing micro air vehicles. Thes e methods achieve high detection performance while remaining light enough for microcontroller and single-board deployment.&nbsp\;</span><span style= "mso-fareast-font-family: 'Times New Roman'\; color: black\;">The second p art outlines the direction for the remaining programme: assessing how resi lient physics-informed detectors are to adversarial evasion crafted within physical constraints\, and developing defences that harden them against a daptive adversaries. The talk closes with the open questions and benchmark s needed to move from detection toward trustworthy\, attack-resilient auto nomy.</span></p> END:VEVENT END:VCALENDAR