Biomedical Continuous-Time Digital Signal Processing Applications
Abstract
Continuous-time digital signal processing (CT-DSP) is an emerging subfield of signal processing. CT-DSP offers three advantages for low power medical electronics and biomedical sensing applications: fewer sample points with higher information content (form of lossless compressive sensing), lower power, and better signal-to-noise ratio compared to conventional digital signal processing. For continuous time sampling, a 3 times reduction in the number of electrocardiogram sample points has been demonstrated while obtaining a 97 accuracy rate in identifying heart arrythmias. A continuous-time control system for a switching DC-to-DC converter has been developed with 3 times improvement in control system signal overshoot. For a 4-bit equivalent analog-to-digital converter (ADC), an offline reconstruction of a continuous-time signal has achieved greater than 100-dB signal-to-noise-and-distortion (SINAD) ratio. A continuous time pipeline ADC has been developed. This ADC overcomes limitations present in conventional pipeline ADCs. A significant issue with continuous-time systems is signal reconstruction. Discrete time systems are linear time-invariant (LTI), and signal reconstruction is conveniently delayed until the final step in the signal processing chain. Continuous-time systems are not LTI, and signal reconstruction is not time-invariant. Research has shown that the simplest CT-DSP reconstruction technique provides modest improvement in SINAD compared to conventional DSP. Initial results from ongoing real-time reconstruction research indicate that it is possible for a 30-dB SINAD improvement. This report presents an introduction to CT-DSP for medical and biomedical sensing applications. The potential improvements covering low power, better signal-to-noise ratio, and fewer data points offer significant capability improvements for low power, battery operated, biomedical applications.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 2024
- Accession Number
- AD1226135
Entities
People
- Patrick Jungwirth
- Sina Najmaei
- W. M. Crowe
Organizations
- United States Army Research Laboratory