Spatio-Temporal Signal Twice-Whitening Algorithms on the hx3100 Ultra-Low Power Multicore Processor
- ORNL
- Coherent Logix, Inc.
- Science Applications International Corporation, Arlington, VA
- Office of Naval Research
While modern signal detection theory fully accounts for spatially distributed sensors, exploiting these techniques for real-time sensing using large, underwater acoustic arrays requires advances in the spatio-temporal signal processing algorithms. In particular, the computational complexity of many spatio-temporal processing techniques is so large that conventional computer processors lack sufficient throughput to provide real-time processing of large spatio-temporal data sets. These limits are exacerbated when constraints, such as power consumption or footprint, reduce the available computational resources. In this report, we demonstrate an implementation of a signal twice-whitening algorithm that is better suited for processing spatio-temporal data in real time. We emphasize these advances by implementing data whitening on the Coherent Logix hx3100 processor, a programmable multicore processor intended for low-power and high-throughput signal processing. These results serve as an example of how the novel capabilities available from emerging multicore processor platforms can provide real-time, software-defined processing of large data sets acquired by spatially distributed sensing.
- Research Organization:
- Oak Ridge National Laboratory (ORNL)
- Sponsoring Organization:
- ORNL work for others
- DOE Contract Number:
- AC05-00OR22725
- OSTI ID:
- 982124
- Country of Publication:
- United States
- Language:
- English
Similar Records
Wavefront array processors - Concept to implementation
Efficient Heterogeneous Execution on Large Multicore and Accelerator Platforms: Case Study Using a Block Tridiagonal Solver