Factor Graph-based Spoofing Mitigation using the Chimera Signal Enhancement

Abstract

To protect civilian GPS signals from spoofing attacks, the Air Force Research Lab (AFRL) has introduced the Chips-Message Robust Authentication (Chimera) signal enhancement for the GPS L1C signal. In the future, users with a compatible receiver can use the Chimera signal enhancement to authenticate GPS measurements every 3 minutes if only using Chimera GPS signals or every 1.5 or 6 seconds with access to a secure out-of-band channel. However, for fast moving vehicles, even the 6 second latency for the fast channel can be high, forcing users to choose between high-latency, authenticated GPS measurements or low-latency, vulnerable GPS measurements. In these cases, it is possible to use proprioceptive sensors, like inertial measurement units (IMUs) and wheel encoders, to validate unauthenticated GPS measurements and lower the latency of authentic position estimates. In this work, we propose using a factor graph optimization (FGO)-based approach to combine IMU, wheel encoder and GPS position measurements with the goal of mitigating spoofing attacks that might occur between Chimera authentications. We use the fault robust method of switchable constraints to introduce a weighting factor for GPS measurements. We use the value of the switchable constraint to indicate whether signals are authentic or not and include a regularization term to indicate that the presence, or absence, of spoofing is likely to be similar across intervals of time. We also introduce the Chimera authentication as a loop closure measurement in the factor graph which fully weights and uses GPS measurements for previous time instances when authentication is received. We experimentally validate our proposed approach in simulations, comparing to a naive FGO implementation, to show that our method successfully mitigates spoofing attacks and utilizes the loop closure to improve position estimates when Chimera authentication is available.