<p>In this paper, we propose conditions for the measurement matrix and signal reconstruction method in single-path indirect time-of-flight (ToF) distance measurement using compressed sensing. Compressed sensing has been widely applied to ToF and used in the spatial domain to improve spatial resolution, and in the time domain to reduce the number of measurements. In indirect ToF distance measurement using compressed sensing in the time domain, measurement matrices based on random, or the Hamming distance of adjacent measurement columns have been proposed. However, there will be the case that in principle, the signals cannot be reconstructed when generating the measurement matrix randomly. In addition, an unrealistic constraint arises that the width of reflected pulses must be equal to the width of the time window. To resolve these issues, we propose a no blind measurement matrix that can reconstruct signals even when the width of the reflected pulse is greater than the width of the time window. Further, the proposed reconstruction algorithm applies orthogonal matching pursuit to single-path indirect ToF distance measurement, thereby enabling fast reconstruction with a finite number of computations. Numerical simulations confirm that under ideal conditions, the proposed method can reliably reconstruct a single pulse and achieve reconstruction in ~ 1/12 of the time compared to that when using iterative reconstruction methods.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Blind-Spot Free Measurement Matrix for Compressed Sensing in Single-Path Indirect Time-of-Flight

  • Keisuke Uchida,
  • Munenori Takumi,
  • Katsuhiro Ishii,
  • Ken-Ichi Kitayama

摘要

In this paper, we propose conditions for the measurement matrix and signal reconstruction method in single-path indirect time-of-flight (ToF) distance measurement using compressed sensing. Compressed sensing has been widely applied to ToF and used in the spatial domain to improve spatial resolution, and in the time domain to reduce the number of measurements. In indirect ToF distance measurement using compressed sensing in the time domain, measurement matrices based on random, or the Hamming distance of adjacent measurement columns have been proposed. However, there will be the case that in principle, the signals cannot be reconstructed when generating the measurement matrix randomly. In addition, an unrealistic constraint arises that the width of reflected pulses must be equal to the width of the time window. To resolve these issues, we propose a no blind measurement matrix that can reconstruct signals even when the width of the reflected pulse is greater than the width of the time window. Further, the proposed reconstruction algorithm applies orthogonal matching pursuit to single-path indirect ToF distance measurement, thereby enabling fast reconstruction with a finite number of computations. Numerical simulations confirm that under ideal conditions, the proposed method can reliably reconstruct a single pulse and achieve reconstruction in ~ 1/12 of the time compared to that when using iterative reconstruction methods.