Processor Architectures for Multi-Mode Multi-Sensor Signal Processing

Toyon Research Corporation proposes to develop a sensor resource manager that will allow an AESA radar to be more successful at using its energy to detect, locate, and identify targets. The sensor manager will utilize a model of the AESA radar and its signal processing functions to select sensor schedules that maximize mission effectiveness. We will work with the Navy to select a system of interest, such as the Fire Scout UAV, which is an ideal candidate in our view. We will work with the developer of the selected AESA radar to quantify the possible modes and waveforms that can be produced by the radar as well as the expected information that can be obtained by the signal processing algorithms. Our algorithms will use this information to compute effective sensor schedules by examining a large number of candidates. The computing resources required to pick the best schedule are potentially enormous for real time operation, thus we will explore the use of different processing architectures for performing our computations. We will analyze the potential use of CPU, GPU, and FPGA processors for this application. In Phase I, we will demonstrate our approach using a CPU, but in a Phase II effort, we will implement our algorithms in the processor selected as a result of the Phase I analysis.