Transcript Slide 1
Thoughts on GPS Security and Integrity Todd Humphreys, UT Austin Aerospace Dept. DHS Visit to UT Radionavigation Lab | March 10, 2011 GPS: The Big Issues Weak GPS Signals Like a 30-Watt lightbulb held 4000 km away GPS does not penetrate well indoors GPS is easy target for jamming GPS is vulnerable to natural interference (e.g., solar radio bursts and ionospheric scintillation) Unauthenticated Civil GPS Signals Civil GPS broadcast “in the clear” Makes civil GPS vulnerable to spoofing Emerging Threat: GPS Jamming Emerging Threat: Civil GPS Spoofing Spoofing and Jamming are Different Threats Spoofing is more difficult & costly Spoofing leaves no trace – victim receiver doesn’t know it’s being spoofed Spoofer typically targets a single receiver Many countermeasures to jamming are ineffective against spoofing Assessing the Spoofing Threat Multi-frequency, multi-system receivers inherently resistant to spoofing Vast majority of GPS receivers in critical applications are single-frequency L1 C/A (easily spoofable) Software radio techniques are game-changer, enabling one to “download” a spoofer Strong financial incentives encourage “complicit spoofing” (spoofing one’s own receiver) Timing receivers used in communications infrastructure are attractive target Civil GPS Spoofing Testbed at UT Austin Spoofer GPS L1 C/A output Software radio platform Output precisely synchronized with authentic signals via feedback Finely adjustable output signal strength Remotely commanded via Internet Defender Vestigial signal defense Data bit latency defense Cryptographic defenses Phase trauma monitoring Dual-frequency tracking Inside the Box Digital attenuator for precise control of output signal power Inside the Box Spoofing signal feedback for precise signal alignment Inside the Box Interface board for remote operation Inside the Box Tracking, data-bit prediction, and synthesis on single DSP Total bill of materials: ~$1,000 Civil Anti-Spoofing Techniques Inspired by Work to Date Data bit latency defense (weak but easy to implement) Multi-antenna defense (patented in 1996; strong against single spoofer; fails against multiple spoofers; requires additional hardware) Vestigial signal defense (work in progress; appears strong) Navigation message authentication (strong, practical, more on this later) Cross-correlation using P(Y) code (pioneered by Lo, refined by Psiaki, very strong but not so practical) Thoughts on the Way Forward for Civil GNSS Authentication More signals means more inherent security, but probably insufficient Some civil cryptographic authentication scheme is likely required “Signal definition inertia is enormous” – Tom Stansell Navigation message authentication (NMA) appears to be best, practical option (advocated by Logan Scott in 2003, others since, more on this later) Goal of cryptographic authentication: force adversary to use directional antennas in a replay attack Preliminary evaluation of NMA for L2C suggests optimism (more on this later) Cryptography must be paired with detection theory Spoofing Detection as a Hypothesis Testing Problem (Soft W-chip Estimation) Spoofing depends on rough See detection forthcoming paper estimates of nominal (C/No)s and (C/No)r on this topic: “Detection strategies for civil cryptographic anti-spoofing.” Navigation and Timing Resilience Through Opportunistic Navigation Tightly-Coupled Opportunistic Navigation Enabling configuration: (1) Same clock: Downmix and sample GPS and SOO with same oscillator (2) Same silicon: Sample GPS and SOO in same A/D converter TCON for Legacy GPS Receivers: The GPS Assimilator Assimilator Prototype More Information http://radionavlab.ae.utexas.edu Backup Slides Synchrophasor-Aided Power Distribution Usage Example: Protecting a GPS Time and Frequency Receiver Usage Example: Reducing Ionospheric Errors Usage Example: Harnessing CDMA Cellular Signals as Aid for Weak GPS Signal Tracking Usage Example: Iridium-Augmented GPS Strong signals Stable clocks Navigational backup to GPS Civilian Anti-spoofing GPS Signals Aiding signal from LEO high-power spot beams over area of operations LEO crosslinks User 400-km switchable beams