A Mapping Methodology for Space-Time Adaptive Processing in Heterogeneous Processors Environment

Abstract

Space-time adaptive processing (STAP) is the key technology of new generation airborne digital signal processing systems. A major challenge in implementing STAP is to meet the real-time requirement while using fewer processors. To meet these requirements, a unified framework for STAP task mapping is developed. In this framework, a STAP task model is firstly brought forward, which exploits the fact that tasks in each stage are independent. Different types of processors are considered and a four-step task mapping methodology is developed. Experimental results show that our methodology can save processors while meeting real-time requirement