Obstacles and Current Approaches for LiDAR Miniaturization In the April 2023 issue of ADAS & Autonomous Vehicles International, Dr. Sabbir Rangwala presents a featured cover article on miniaturization of LiDAR sensors.

Highlights

  1. Obstacles to Lidar Miniaturization:
    • Challenges in laser production efficiency and transmission in free space.
    • Inefficient transmit-receive process and eye safety concerns.
    • Laser inefficiency and sensitivity to temperature, requiring complex packaging.
  2. Current Approaches:
    • Solid-state methods for Field of View (FoV) without moving parts.
    • Use of single laser pulses or electronic scanning arrays for addressing FoV.
    • Utilizing advancements in semiconductor technology for lidar development.
    • Mention of specific companies and their lidar technologies.

Obstacles to lidar miniaturization

The main obstacle to making LiDAR smaller lies in the laser technology it utilizes. Generating light particles, known as photons, from electrons is a complex and inefficient process. In the 1990s, the telecommunications revolution played a crucial role in advancing semiconductor lasers from research laboratories to mass production facilities and integrating them into both land-based and undersea fiber-optic networks. However, LiDAR faces unique challenges due to the need for transmitting laser energy through free space.

The process of sending and receiving laser signals in LiDAR is inefficient because it is affected by atmospheric absorption and losses in the optical connection related to the distance of transmission. Achieving high-resolution imagery and fast frame rates over a substantial field of view (FoV) demands power levels that semiconductor lasers struggle to provide. This has led to the utilization of techniques such as optical amplification (using fiber lasers), employing large laser arrays (like VCSELs), or sharing laser energy in both time and space (through scanning mechanisms).

Safety is a significant concern, especially concerning human eyes. Some LiDAR systems use wavelengths in the range of 800-900nm, which have limited safety margins for the eyes. The safety improves with the use of 1,300-1,500nm lasers, but there are still restrictions on the maximum safe power density to maintain a certain performance level. Designing eye-safe solutions necessitates bulky system packaging and specialized optics.

Laser systems are known for their inefficiency and sensitivity to temperature. A substantial portion (approximately 70-80%) of the electrical energy utilized in lasers gets converted into heat, which requires effective management strategies. Automotive temperature variations present additional challenges, causing shifts in the laser wavelength and further degradation of efficiency. The III-V semiconductors (like GaAs or InGaAs) commonly used in lasers degrade faster at higher temperatures and in the presence of moisture. To counter these issues, active cooling and more intricate packaging solutions are vital.

In the broader context of a LiDAR system, successful miniaturization requires a hybrid integration approach utilizing diverse materials: complex III-V semiconductors, silicon-based electronic components, glass fibers, bulk optics (such as focusing lenses and isolators), scanning mechanisms, effective thermal management, and sophisticated packaging methods.

Current Approaches

The field of view (FoV) in LiDAR technology is currently being tackled through solid-state methods, which do not involve moving parts. There are two main approaches to achieving this:

  1. Single Laser Pulse or Flash: In this method, a single laser pulse or flash is used to simultaneously address all image pixels. Some notable companies employing this approach include PreAct, TriEye, and Ouster.
  2. Electronic Scanning Arrays: This approach uses electronic scanning arrays composed of monolithic silicon SPADs (Single-Photon Avalanche Diodes) and GaAs VCSELs (Vertical-Cavity Surface-Emitting Lasers) to sequentially address different regions within the FoV. Companies like Opsys and Hesai utilize this technology.

The VCSEL-SPAD approach benefits from advancements, commercialization, and integration of Time-of-Flight (ToF) LiDAR in smartphones, typically operating at 905/940nm wavelengths (exact values may vary and are proprietary). Another technique involves optical scanning through a combination of phase-tuning antennas known as optical phase arrays (OPAs) and wavelength dispersion. This is implemented in chip-scale silicon photonics platforms, with Analog Photonics being a notable player in this field. This platform is compatible with Frequency-Modulated Continuous Wave (FMCW) coherent LiDAR, which simultaneously measures range and radial velocity and operates in the 1,500nm wavelength band.

PreAct specializes in short-range LiDAR for in-cabin sensing and road-facing applications. Their approach is innovative, using low-cost, readily available CCD arrays and LED light sources (rather than lasers) to generate 3D images based on indirect time of flight (iToF) principles, similar to those used in gaming applications. Their TrueSense T30 LiDAR impressively operates at a high frame rate of 150Hz, crucial for quick responses in short-range applications such as blind spot obstacle avoidance and pedestrian safety. The device’s size envelope includes an 8MP RGB camera and electronics that combine visible and 3D images. By eliminating the RGB sensor, the size can be further reduced.

TriEye’s SEDAR (Spectrum Enhanced Detection and Ranging) is a flash LiDAR system that employs a 1.3Mp CMOS-based germanium-silicon SWIR (Short-Wave Infrared) detector array and an internally developed, Q-switched, high peak-power, solid-state pumped diode laser to illuminate the entire FoV. The use of a higher wavelength enhances eye safety margins, allowing for the utilization of higher laser power.

Opsys has a distinctive approach, utilizing electronically addressable high-power VCSEL and SPAD arrays to create a solid-state LiDAR with no moving parts. This system can operate across automotive temperature ranges without the need for active cooling or temperature stabilization.

Hesai is actively producing long-range LiDAR systems for multiple automotive customers with the AT128 model, which employs mechanical scanning for a high FoV. They also offer the FT120, a fully solid-state LiDAR that uses electronic scanning of VCSEL and SPAD arrays and is designed for short-range applications like blind spot detection and in-cabin use. Hesai became a publicly traded company in January 2023 and is currently in a quiet period, suggesting ongoing developments in their LiDAR technology.