Abstract
The probability of detection in radar systems depends on interconnected parameters that may be disrupted under adverse conditions such as heavy rain, clutter, or jamming. This paper explores the applicability of resilience models that incorporate regression, time series, and mixture methods to predict radar performance under such conditions. A simulated radar scenario with varying internal parameters and heavy rain was used to evaluate the model performance. Results demonstrate that the hybrid combination between regression and time series components achieve superior accuracy by jointly capturing current effects and temporal correlations, enabling accurate performance prediction with 90% empirical coverage and reliable estimation of resilience metrics. These predictive capabilities support proactive management by identifying vulnerabilities and guiding recovery strategies, thereby strengthening the resilience of radar systems.