ANALYSIS AND DETECTION OF GPS SPOOFING ATTACKS IN CYBER PHYSICAL SYSTEMS

Abstract

GPS is relied upon by most cyber-physical systems because of its ability to provide a precise time or position reference. The integrity and availability of GPS signals have a significant impact on the safety and security of cyber-physical systems. This dissertation presented a detailed analysis of GPS time spoofing attacks and the effectiveness of different spoofing methods. And two light-weight detection techniques are proposed to mitigate the impact of general GPS spoofing attacks. One detection technique is based on the auto-correlation among consecutive GPS signals, where runs test is used to quantify the correlation and supervised learning techniques are used to classify. Another detection technique is based on the channel characteristics of the GPS signal, which performs dynamic threshold detection on the established relationship between the signal envelope characteristics and the signal propagation distance. Validated by experiments using actual GPS devices, both proposed techniques are efficient and have high accuracy.