A method is developed that processes GNSS beat carrier-phase measurements from a single moving antenna in order to determine whether the GNSS signals are being spoofed. This technique allows a specially equipped GNSS receiver to detect sophisticated spoofing attacks that cannot be detected using standard RAIM techniques. The new method does not require changes to the GNSS signal structure. It can detect spoofing attacks on unencrypted civilian signals as well as meaconing attacks on encrypted military signals. The method uses short segments of beat carrier-phase time histories from multiple satellites that are collected while the receiver's single antenna is undergoing a known, high-frequency, oscillatory motion profile. Normally this profile would be pre-programmed into an antenna articulation system. The antenna also can be moving in an unknown way at lower frequencies, as might be the case if it were mounted on a ground vehicle, a ship, an airplane, or a spacecraft. The spoofing detection algorithm correlates high-pass-filtered versions of the known motion component with high-pass-filtered versions of the carrier-phase variations. True signals produce a specific correlation pattern, and spoofed signals produce a recognizably different pattern if the spoofer transmits its false signals from a single antenna. The most pronounced difference is that multiple non-spoofed signals display variations between their beat carrier-phase responses, while all signals' responses are identical in the spoofed case. These differing carrier-phase profiles have been used to develop a precise hypothesis test to detect whether or not a spoofing attack is in progress. The test statistic is the difference between the mean-square fit errors of two carrier-phase models, one a spoofed model that expects identical carrier-phase oscillations and the other a non-spoofed model that expects differing oscillations on the different channels. For moving-base receivers, there is no need for prior knowledge of the vehicle's attitude; the detection synthesizes an attitude estimate as part of its calculations for the non-spoofed hypothesis. A prototype version of this spoofing detection system has been designed and tested against actual live-signal spoofing attacks on the GPS L1 C/A code. The spoofer used to generate the attacks was an advanced version of the device described in Ref. 1. The live-signal tests were arranged by the U.S. Department of Homeland Security. They were conducted in June 2012 at the White Sands Missile Range, NM under the supervision of the U.S. Air Force 746th Test Squadron. The prototype system correctly identified spoofing in the 4 cases out of 8 live-data trials that involved actual attacks. These detections used worst-case false-alarm probabilities of 1.e-06, and their worst-case probabilities of missed detection were no greater than 1.6e-06. The ranges of antenna motion used to detect spoofing in these trials were between 4 and 6 cm, i.e., on the order of a quarter-cycle of the GPS L1 carrier wavelength. The prototype system recorded digitized RF data and performed its signal tracking and spoofing detection calculations using off-line software radio techniques. The data spans used for spoofing detection were about 0.5 seconds long. The detection calculations could easily be implemented in real-time. Given sufficiently fast antenna oscillations, they could operate on carrier-phase data spans of 0.125 sec or less, and they could achieve probabilities of missed detection and false alarm that were both on the order of 1.e-06. References: [1] Humphreys, T.E., Ledvina, B.M., Psiaki, M.L., O’ Hanlon, B.W., and Kintner, P.M. Jr., “Assessing the …