Multistage algorithms for extended dwell target detection

Improvements in detecting weak targets from a small radar platform must come through increased temporal integration, i.e., extending the time over which target samples are coherently integrated. Conventional single-channel radar assumes a linear-phase signal model that is only accurate over a short dwell time for typical target motion. Over an extended dwell, the target signal includes multiple nonlinear phase components, each of whose effects become significant at different times during the dwell. An algorithm is presented that develops a multiphase signal model in multiple stages based on these times. A modification of the proposed approach improves the signal model for the most challenging targets. When used as the detection filter, the multiphase signal model yields near optimal performance over an extended dwell time for a wide range of target parameters. Typical improvement in output signal-to-noise ratio (SNR) for a 500 ms dwell is 12-13 dB over conventional processing.