Abstract:

The use of terahertz (THz) wave for absolute positioning has recently been demonstrated in [1]–[4] for the sake of contactless sensing, operation in adversarial conditions (e.g., fire, smoke), and robustness to dust and dirt. THz absolute positioning can operate in a raster [1], [2] or compressed scanning mode [3], [4]. In the raster scanning mode of Fig. 1 (a), the scale with THz-band QR patterns is illuminated by a single THz transceiver, and a programmable mechanical raster moves the transceiver in one or two dimensions. In the compressed scanning mode of Fig. 1 (b), the THz positioning system also uses a single THz transceiver but with random masks and collimating/focusing lens to cover a large area (versus a single pixel in the raster scanning) of the scale encoded by QR patterns which can be uniquely mapped to positions. Consequently, signal processing algorithms are required to detangle the compressed THz signal and recover the underlying QR pattern due to the mixing between random masks and QR patterns. In [3], [4], we proposed a variational Bayesian inference (VBI) method for this task by exploiting the non-negative and binary features of THz reflectance. This paper aims to report the first experimental result to verify the THz QR positioning in the compressed scanning mode.