Adaptive detection in dense target environments

We consider the impact of target signals corrupting the covariance estimate when implementing STAP in ground moving target indication (GMTI) scenarios. Herein, we propose a model for seeding the radar data cube with target signals and calculate both asymptotic and finite training data detection losses. In general, both the location and strength of the corruptive target signals influence performance: under practical conditions, SINR loss can exceed 10 dB. Measured MCARM data confirms the presence of ample ground moving target signals in the radar field of regard under typical operating conditions. We conclude by briefly considering several mitigating strategies as well.