|
Our research program aims to advance green embedded computing technologies targeting biomedical and industrial applications of ultrafast ultrasound imaging. Such imaging, based on the use of unfocused acoustic beams, is characterized by very high rates of raw data acquisition, which is needed for accurately capturing dynamic biomedical phenomena (e.g., shear waves in tissue) or for quickly collecting large volumes of nondestructive testing data (e.g., inline inspection of pipelines).
Efficient data acquisition, probe-to-processor data transfer, and subsequent data processing (i.e., image reconstruction, enhancement, and analysis) require innovative algorithms and architectures to meet application demands in a tightly constrained embedded computing environment. Operational constraints include not only communication, computation, and memory limitations, but also a stringent energy budget imposed by finite-capacity batteries powering untethered embedded systems.
Our current research objectives fall into the following main categories:
- Development of innovative low-complexity algorithms for solving multidimensional data processing problems that arise in ultrafast ultrasound imaging applications;
- Development of innovative configurable hardware/software architectures for supporting energy-efficient execution of these algorithms;
- Development of interpretable predictive models producing battery-aware system-level energy budget allocations to guide energy-constrained optimizations of application performance.
|
|
