Loading...

Iterative Methods for Sparse Reconstruction in Level Crossing Analog to Digital Converters

Boloursaz Mashhadi, Mahdi | 2018

1596 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 51146 (05)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Marvasti, Farokh
  7. Abstract:
  8. In this research, we propose analog to digital (A/D) converters based on Level Crossing (LC) sampling and the corresponding signal processing techniques for effecient acquisition of spectrum-sparse signals. Spectrum-sparse signals arise in many applications such as cognitive radio networks, frequency hopping communications, radar/sonar imaging systems, musical audio signals and many more. In such cases, the signal components maybe sparsely spread over a wide spectrum and need to be acquired at a reasonable cost without prior knowledge of their frequencies. Compared with the literature, the proposed scheme not only enables efficient acquisition of spectrum-sparse signals with a less complex sampling structure, but also significantly decreases power consumption in the circuitry by enabling asynchronous implementation. In this research, we introduce the basic LC constraint and propose algorithms for sparse signal reconstruction from LC samples based on iterative thresholding and convex projection techniques. Utilizing the proposed algorithms, we show improved performance of the proposed A/D conversion scheme in comparison with the classic A/D converters based on low pass signal assumption for audio signals. Furthermore, we propose two algorithms for sparse reconstruction from LC samples utilizing tools from the theory of one-bit Compressive Sensing (CS) to enforce the basic LC sampling constraint. We also propose a novel feedback acquisition scheme for efficient sampling of spectrum-sparse signals utilizing level comparisons. In the proposed scheme, the spectral components are estimated utilizing a sparsity-promoting, sliding-window algorithm in a feedback loop. Utilizing the estimated spectral components, a level signal is predicted and sign measurements of the prediction error are acquired. Unlike many batch-based compressive sensing algorithms, our proposed algorithm gradually estimates and follows slow changes in the sparse components utilizing a sliding-window technique. We also consider the challenges to practical implementation of the proposed scheme and propose techniques to cope with propagation of the sign-flip errors in the feedback loop
  9. Keywords:
  10. Adaptive Level-Crossing Sampling ; Analog to Digital Converter ; Compressive Sensing ; Sparse Signal Reconstruction ; One-Bit Compressive Sensing

 Digital Object List

 Bookmark

...see more