Bias detection and its confidence assessment in Global Positioning System signals

The Federal Aviation Administration plans to migrate to a Global Positioning System (GPS) based navigation system for aeronautical use. As with any safety-critical application, any system failure mode, which might lead to misleading information (MI) for the user, needs to be detected at a high level of confidence. However, the failure detection schemes cannot be too conservative in order not to unreasonably decrease the system's continuity and/or availability. This work focuses on the detection of GPS signal-in-space anomalies. Since the received GPS signal is altered by noise, any signal parameter determination translates into a statistical estimation process. An estimation procedure to analyze the received satellite signal for possible biases is introduced in this paper. Since the statistical behavior of an anomalous signal is unknown, a distribution-independent detection scheme is presented. Further, the algorithm also provides a confidence assessment on the statistical inference. The derived bias detection procedure is corroborated with simulation results.