Advanced AI for Modern C-UAS Operations

At CoVar, we recognize that counter-UAS requires a layered approach rather than a single solution. We deliver a suite of AI-enabled capabilities that address operational needs from long-range detection to close-range classification and payload disambiguation. At extended ranges, where low pixels-on-target limit traditional image-based detectors, CoVar leverages UAS motion cues through a spatio-temporal deep network to enable reliable long-range detection. This approach outperforms both image-only methods and statistical moving-object-indicator models, providing earlier warning and increased confidence at standoff distances.

At closer ranges, our approach enables target classification and payload disambiguation by estimating a 3D point cloud of the UAS from passive sensor imagery, allowing operators to maintain electromagnetic concealment. This method increases interpretability and supports zero-shot classification by comparing point clouds against a library of CAD models, enabling rapid integration of new targets without retraining. Once classified, CoVar can disambiguate payloads, such as distinguishing cameras from munitions, supporting real-time risk assessment, target prioritization, and informed engagement decisions.