Quantum Induced Coherence Light Detection and Ranging

Quantum illumination has been proposed and demonstrated to improve the signal-to-noise ratio (SNR) in light detection and ranging (LiDAR). When relying on coincidence detection alone, such a quantum LiDAR is limited by the timing jitter of the detector and suffers from jamming noise. Inspired by the Zou-Wang-Mandel experiment, we design, construct, and validate a quantum induced coherence (QuIC) LiDAR which is inherently immune to ambient and jamming noises. In traditional LiDAR the direct detection of the reflected probe photons suffers from deteriorating SNR for increasing background noise. In QuIC LiDAR we circumvent this obstacle by only detecting the entangled reference photons, whose single-photon interference fringes are used to obtain the distance of the object, while the reflected probe photons are used to erase path information of the reference photons. In consequence, the noise accompanying the reflected probe light has no effect on the detected signal. We demonstrate such noise resilience with both LED and laser light to mimic the background and jamming noise. The proposed method paves a new way of battling noise in precise quantum electromagnetic sensing and ranging.

Figure 1

Physical principles for conventional, QI and QuIC LiDARs. (a) Conventional LiDAR. (b) QI LiDAR, where the SNR is enhanced by joint measurement. (c) QuIC LiDAR. Without detecting the probe light, a locally kept reference light is detected to obtain the distance of the object, enabled by QuIC. (d) The setup consists of three modules (enclosed in dashed lines), the pump laser, the local scanning module and the detection module. A collimated 532 nm laser (green line, 10 kHz linewidth) is used to pump a type-0 PPLN crystal (2 cm length) to generate SPDC entangled reference (blue line) and probe (yellow line) photons centered at 893 and 1316 nm wavelength. The two light beams are separated by a dichroic mirror ${\mathrm{DM}}_2$. The reference beam is reflected by a scanning mirror ${\mathrm{M}}_r$. All reflected beams go through the crystal a second time. We detect the reference beam with a Si-based CMOS camera. The probe light is also detected with an InGaAs camera to compare the performance of QuIC and conventional LiDAR. An off-axis parabolic mirror and plano-convex lenses are used for the imaging setup (see [22] for details).

Figure 2

Ranging of QuIC LiDAR. (a) Real-time images captured by the CMOS camera during the scanning of ${\mathrm{M}}_r$. The two images carrying interference patterns correspond to the two faces of the plate. The exposure time of each image is 50 ms. The scanning step is 60 nm with a total scanning length 3.5 mm. (b) Interference fringes along the white lines in (a). The inset shows the interference fringes in a region within the red rectangle. (c) Interference visibility obtained from STFT with an integration window $100\mathrm{\mu }\mathrm{m}$. The step in the horizontal axis is $1\mathrm{\mu }\mathrm{m}$. Inset: the Fourier transform of the interference in the red frame in (b). The Fourier transform intensity in the spatial frequency range from 1.5 to $3{\mathrm{\mu }\mathrm{m}}^{-1}$ (blue shaded area) is used to extract the visibility (see [22] for detail). The two peaks in the main frame with the weights being marked by red dashed lines indicate the positions of the two surfaces. (d) Images and positions of the two silica plate faces obtained from QuIC LiDAR. We plot the visibility as functions of the positions of the STFT weights recorded on the reference camera during the delay scan. The thickness of the silica plate is $1.01\pm 0.01\mathrm{mm}$. The optical path between the two faces measured from QuIC LiDAR is $1.499\pm 0.009\mathrm{mm}$.

Figure 3

Noise robustness of QuIC LiDAR. The images in the upper (lower) row are extracted from the reference (probe) light cameras. (a) and (b) are images without added noise. (c) and (d) are images in LED noise. (e)–(h) are images in jamming laser noise. (e) and (f) satisfy the phase matching condition while (g) and (h) do not. The powers of the reflected probe beam, the LED light and the laser before entering the probe camera are 72 nW, 44 nW, and $2.6\mathrm{\mu }\mathrm{W}$. The exposure time is 6 ms for (a), (c), (e) and (g), 15 ms for (b) and (d), and $20\mathrm{\mu }\mathrm{s}$ for (f) and (h). Here, (f) and (h) are unprocessed images from the probe camera. The reference image has lower resolution than the probe image due to the imperfection of the transverse phase matching in the crystal, which is restricted by the crystal aperture and the pump beam waist size [51].

Figure 4

The SNR of QuIC LiDAR in LED (a) and laser (b) noises. In (a), we compare the SNR in QuIC LiDAR (blue triangles) and in direct detection of the probe light (red circles). For noise level below 23.4 dB, the LED noise is weaker than the systematic noise, such that the SNR of the probe light remains unchanged. For higher LED noise, the SNR of the probe light decreases, while the one of the QuIC LiDAR remains unchanged. In (b), we simulate saturation jamming by sending a laser back to the crystal, which induces stimulated PDC noise. The red circles and blue triangles are the SNRs of the reference light in the areas satisfying and violating the phase matching condition. The intensities of probe and noise light are measured by an InGaAs power meter with an iris covering the region of interest on the camera.