Study Focused on Photonics Circuitry Shows Promising Advances for the Use of LiDAR in Autonomous VehiclesMay 26, 2020 by Luke James
Researchers at the lab of Tobias Kippenberg at Ecole Polytechnique Federale de Lausanne (EPFL) have reportedly found a new way to implement a parallel FMCW LiDAR engine by using integrated nonlinear photonic circuitry.
Coherent ranging, also known as frequency-modulated continuous-wave (FMCW) laser-based LiDAR is used for long-range three-dimensional distance and velocimetry in autonomous driving.
Using frequency-chirped waveforms, FMCW LiDAR maps distance to frequency and simultaneously measures the Doppler shift of reflected laser light, similar to sonar or radar, and coherent detection prevents interference from sunlight and other LiDAR systems.
However, the technical complexity of precisely controlling narrow-linewidth frequency-agile lasers has until now prevented the successful parallelization of FMCW LiDAR. Now, researchers have reportedly demonstrated a massively parallel coherent LiDAR scheme using an ultra-low-loss photonic chip-based soliton microcomb.
Converting Continuous Wave Laser Light Into a Stable Optical Pulse
The research team coupled a single FMCW laser into a silicon-nitride planar microresonator. Here, the continuous wave laser light is converted into a stable optical pulse due to the double balance of dispersion, nonlinearity, cavity pumping, and loss.
"Surprisingly, the formation of the dissipative Kerr soliton, does not only persist when the pump laser is chirped, but transfers the chirp faithfully to all the generated comb teeth," says Johann Riemensberger, postdoc at Kippenberg's lab and first author of the study.
A graphical representation of LiDAR waves. Image used courtesy of Johann Rimensberger of EPFL
Improving FMCW LiDAR ‘Tenfold’
The microresonator’s small size means that the comb teeth are spaced 100GHz apart. This is enough to separate them by using standard diffraction optics, and because each comb tooth inherits the linear chirping of the pump laser, the team discovered that it was possible to create up to 30 independent FMCW LiDAR channels per microresonator.
Each channel in turn can simultaneously measure distance and velocity while the spectral separation of the channels makes the device immune to channel crosstalk, as well as inherently natural for co-integration with recent optical phased arrays based on photonic integrated optical grating emitters.
Furthermore, the spatial separation of emitted beams and 1500 nm-wavelength operation relaxes stringent eye and camera safety limitations. "The technology developed here at EPFL could improve acquisition rates of coherent FMCW LiDAR tenfold in the near future," says Anton Lukashchuk, PhD student. The team’s concept rests on high-quality silicon-nitride microresonators that exhibit record-low losses among planar nonlinear waveguide platforms. These were produced at EPFL’s Centre of MicroNanotechnology (CMi) and are already commercially available.
This work may pave the way for the widespread adoption of coherent LiDAR in tomorrow’s autonomous vehicles, making them safer and better at detecting obstacles in their environments.