Analyzing the Impact of Sensor Noise in the TCPA-DCPA Domain
Aniek De Booij, Erik Theunissen · 2024
To provide timely, consistent alerts in the operational environment, the alerting function of a Detect and Avoid system must be sufficiently robust to the impact of measurement uncertainty. To evaluate the performance of Sensor Uncertainty Management (SUM) using simulation, this uncertainty is typically modelled as a certain type of noise which is added to the truth data of a conflict geometry. The resulting noisy trajectories are provided as input to the alerting function. The Technical Performance Metrics include the number of missed alerts, false alerts, late alerts and intermittent alerts. For evaluation purposes, it is important to be able to distinguish between failures that can likely be prevented by improving SUM, and failures due to conflict geometry/sensor combinations that are highly susceptible to sensor noise. In this paper, a method is described that supports a classification of the input used for testing based on the likelihood of sensor noise causing the predicted miss distance to exceed worst-case thresholds.