Three reasons why MEMS LiDAR is considered the fastest landing technology

MEMS LiDAR has always been regarded as the fastest commercial LiDAR technology route to land in the field of autonomous driving. Only a quarter of 2019 has passed, and the news of investment in the MEMS LiDAR field and the launch of various new products make us strongly feel the footsteps of its landing getting closer and closer!


In just March, MEMS LiDAR manufacturer Innoviz Technologies (partnering with BMW to integrate MEMS LiDAR into cars in 2021) announced the completion of a Series C financing of a total of $132 million. In addition to Israeli investment institutions, Chinese investment institutions (China Merchants Capital, Shenzhen Venture Capital, and Lianxin Capital) have also emerged as investors. At the CES 2019 held in Las Vegas, USA in January this year, Chinese laser radar leaders Sagitar Juchuang and Hesai Technology respectively launched their own MEMS laser radars: RS-LiDAR-M1 and PandarGT 3.0. Prior to this, Sagitar Juchuang and Hesai Technology were leaders in the field of mechanical LiDAR technology. In the process of evolution from mechanical LiDAR to solid-state LiDAR, some companies choose to directly enter all solid-state LiDAR, while many companies are deeply involved in the hybrid solid-state technology route - MEMS LiDAR. So, will 2019 really become the first year of the MEMS LiDAR technology roadmap?

From Yole's latest report "Lidar for Automotive and Industrial Applications -2019 Edition", it can be seen that MEMS and Flash technology routes are more favored by lidar manufacturers

We know that mechanical LiDAR is bulky and expensive, such as Velodyne's 32 line LiDAR HDL-32E, which requires 32 sets of transmitting light sources and 32 sets of receiving light sources to be adjusted one by one, requiring very high assembly requirements and worrying production and shipping efficiency; Alternatively, a rotating mirror can be used to control a single pulse laser at slightly different tilt angles and orientations, such as the Fareo SCALA. As one of the solid-state LiDARs, Optical Phased Array (OPA) LiDAR has significantly reduced volume, controllable assembly time, and high reliability. However, due to various issues such as insufficient chip maturity, there is still a long way to go before it can be implemented. Flash LiDAR is currently unable to meet the requirements of both far and near imaging, but with the maturity of single photon array detection technology, it is expected to become the future direction of LiDAR technology.

MEMS micro mirrors, also known as MEMS scanning mirrors or MEMS micro mirrors, are uniformly expressed in this article using MEMS micro mirrors. According to principles, it mainly includes four types: electrostatic drive, electromagnetic drive, electric heating drive, and piezoelectric drive. The first two technologies are relatively mature and have wider applications. Texas Instruments (TI) successfully commercialized the electrostatic driven MEMS micro mirror in 1996. 
Schematic diagram of MEMS micro mirror operation


What is MEMS LiDAR? This article refers to the use of semiconductor 'micro motion' devices - MEMS micro mirrors (instead of macroscopic mechanical scanners) to achieve beam manipulation at the micro scale of the lidar emitter as a 'hybrid solid-state'. Meanwhile, the laser detection and ranging system using the aforementioned beam manipulation method is referred to as a hybrid solid-state LiDAR or MEMS LiDAR. So, why did the concept of "mixed solids" arise? Because MEMS micro mirrors are silicon-based semiconductor components that belong to solid-state electronic components; However, MEMS micro mirrors are not "safe" and integrate "movable" micro mirror surfaces internally; From this, it can be seen that MEMS micro mirrors have both "solid" and "motion" properties, hence they are called "hybrid solid". It can be said that MEMS micro mirrors are innovators of traditional mechanical LiDAR, leading the miniaturization and low-cost development of LiDAR. 
Working principle diagram of MEMS LiDAR

The reason why the industry regards MEMS LiDAR as the fastest technological route to land is mainly due to three aspects:

Firstly, MEMS micro mirrors help LiDAR overcome bulky mechanical motion devices such as motors and prisms. Millimeter sized micro mirrors greatly reduce the size of LiDAR. Whether in terms of aesthetics, vehicle integration, or cost, their advantages are amazing! 

Secondly, the introduction of MEMS micro mirrors can reduce the number of lasers and detectors, greatly reducing costs. Traditional mechanical LiDAR requires as many sets of transmitting and receiving modules as needed to achieve as many wire harnesses. By using a two-dimensional MEMS micro mirror, only one laser light source is needed. The laser beam is reflected through one MEMS micro mirror, and the two work together at a microsecond frequency. After being received by the detector, the goal of 3D scanning the target object is achieved. Compared with mechanical LiDAR structures with multiple sets of transmitting/receiving chipsets, MEMS LiDAR significantly reduces the demand for the number of lasers and detectors. From a cost perspective, an N-line mechanical LiDAR requires N sets of IC chipsets: cross impedance amplifier (TIA), low noise amplifier (LNA), comparator, analog-to-digital converter (ADC), etc. Maims Consulting estimates that the cost of each group of chips is about $200, and the cost of only 16 sets of chips is as high as $3200. Innoluce has released a MEMS LiDAR design solution that utilizes MEMS micro mirrors and integrates various discrete chips into the LiDAR control chipset, resulting in the cost of LiDAR being controlled within $200.


Innoluce adopts a MEMS LiDAR design scheme with MEMS micro mirrors, with a cost of less than $200

Thirdly, MEMS micro mirrors are not devices born for LiDAR, they have been commercially applied in the field of projection displays for many years. The most successful application case is Texas Instruments (TI) DLP (Digital) Light  Processing， The core technology of digital light processing (DLP) display is Texas Instruments' unique "black technology" - an array of MEMS micro mirrors based on electrostatic principles, with each micro mirror forming a monochromatic pixel. The register in the lower layer of the micro mirror controls the high-speed switching of specific lenses between on and off states, blending pixels of different colors together. In addition, MEMS micro resonators have also been successfully applied in fields such as 3D cameras, barcode scanning, laser printers, medical imaging, and optical communication.

Up to now, there is only one type of laser radar that is truly vehicle size class, that is, the mechanical laser radar SCALA from Valeo, configured in the Level 3 autonomous vehicle Audi A8 released in 2017. SCALA adopts the Direct Time of Flight (DToF) ranging method, with a rotating scanning mirror as the beam operating unit, a high-power laser diode as the light source, and an avalanche photodiode (APD) array with three sensitive units as the detector. Of course, Fareo will also plan to launch a LiDAR using MEMS micro mirrors: SCALA 3. So, why is MEMS LiDAR full of hope, and MEMS micro mirror technology has matured in other application fields, but there has not yet been a truly automotive grade MEMS LiDAR?


Firstly, in terms of MEMS micro mirrors themselves, the technical threshold is very high. Texas Instruments' DLP technology stands out among the crowd, and the story behind it is that this technology was first introduced in 1987 and was initially only used for defense. It was not until 1996 that it was commercialized. It took a full nine years for this financially strong and technologically developed company to succeed. The difficulty is evident. Mature and mass-produced MEMS micro mirror enterprises are mainly concentrated abroad, such as Innoluce acquired by Infineon in Germany, Mirrorcle in the United States, Hamamatsu in Japan, STMicroVision in Switzerland, and MicroVision in the United States. Fortunately, Chinese MEMS micro mirror enterprises have developed rapidly in recent years, such as Xi'an Zhiwei Sensing, Taiwan Opus, Suzhou Xijing Technology, etc.

Secondly, the success of MEMS micro mirrors in fields such as projection display cannot be replicated in vehicle mounted LiDAR. MEMS micro mirrors belong to vibration sensitive devices, and the vibration and impact in the vehicle environment can easily have adverse effects on their service life and working stability, leading to a deterioration of the measurement performance of LiDAR. Therefore, it is necessary to conduct in-depth research on the isolation vibration technology of MEMS micro mirrors. As a crucial sensor for human life, LiDAR needs to meet both vehicle regulations and mass production. To overcome this gap, it still requires technological advancement and time accumulation.

Again, compared to the prisms and pendulums used in mechanical LiDAR, the size of MEMS micro mirrors has indeed been greatly reduced, but the problem it brings is that it limits the optical aperture and scanning angle of MEMS LiDAR, and the field of view angle will also decrease.


In order to achieve maximum optical aperture, MEMS LiDAR manufacturers pursue large-sized MEMS mirrors. But practitioners in integrated circuit manufacturing know that the larger the chip size, the higher the cost; At the same time, the more sensitive to defects, the yield of chips manufactured on the same wafer is inversely proportional to the size of a single chip, which greatly increases manufacturing difficulty and cost. At the same time, the problem brought about by the large size is the decrease in scanning frequency, which may not meet the requirements of real-time ranging and imaging of in vehicle LiDAR. MEMS LiDAR designers will inevitably face the challenge of balancing size and frequency.

Meanwhile, in order to obtain a larger scanning angle, a MEMS micro mirror with a large deflection angle is required. However, the resolution of the scanning system is determined by the product of the mirror size and the maximum deflection angle, and the mirror size and deflection angle are a pair of irreconcilable enemies. There are two directions to solve this problem: (1) By modulating the driving voltage frequency, the MEMS micro mirror is placed in a resonant working state, and the maximum polarization angle will be amplified; (2) Expand the beam and amplify the maximum polarization angle through optical components such as lenses, diffractive optical elements, and liquid crystal spatial modulators. However, beam expansion will bring numerous complex technical issues, which will not be discussed here. 

Compared with mechanical LiDAR (left), MEMS LiDAR (center), and OPA LiDAR (right) scanning methods, the scanning angle of MEMS LiDAR is limited by the mirror size and deflection angle of MEMS micro mirrors, resulting in a smaller scanning angle

At present, American MEMS micro mirror manufacturer Mirrorcle uses bonding methods to assemble additional large mirror surfaces on the driver after processing, improving the filling ratio. Therefore, it can provide MEMS mirror surfaces with sizes as large as 7.5mm, which is favored by many MEMS LiDAR system manufacturers. However, the price of MEMS micro mirrors with large mirror surfaces in Mirrorcle ranges from several thousand yuan. As a demo product in the early stage (DEMO), I can bear to grit my teeth, but once it's mass-produced, such high costs are not commercially viable. In this situation, we have seen some domestic and foreign laser radar industry chain manufacturers grasp the lifeline of MEMS laser radar through self research or investment/acquisition of companies. If Infineon acquires Innoluce from the Netherlands and provides chips and solutions for MEMS LiDAR manufacturers; Sagitar Juchuang invested in Xijing Technology to lay out MEMS LiDAR. According to previous reports from Maims Consulting, Xijing Technology has developed a MEMS micro mirror with a diameter of 5mm, which has entered the mass production stage; The MEMS micro mirrors used in the PandarGT 3.0 of Hesai Technology are provided by the self-developed team. 

The quotation for MEMS micro galvanometer products provided on the Mirrorcle official website


The working temperature range is also a major threshold for MEMS micro mirrors to pass through vehicle regulations. Normally, automotive grade products require core components to meet the operating range of -40 ℃ to 125 ℃. In practical applications, the material properties of MEMS micro mirrors (such as Young's modulus and shear modulus) will change with changes in environmental temperature, leading to changes in the motion characteristics of micro mirrors. Therefore, the selection of materials and manufacturing processes is a huge challenge for achieving automotive grade MEMS micro mirrors. 

Due to the limited mirror size of MEMS micro mirrors, the receiving aperture of MEMS LiDAR at the receiving end is very small, which has become a thorny issue on its mass production path. Here is some additional knowledge about the LiDAR receiver. Due to the fact that only a small portion of the photons emitted by the pulse can reach the effective area of the receiving photodetector. If the atmospheric attenuation does not change along the pulse path, the divergence of the laser beam can be ignored. When the spot size is smaller than the target object, the incident angle is perpendicular to the detector, and the reflector is a Lambertian body (reflecting in all directions), the peak power P (R) of the light reception is

Among them, P0 is the peak power of the emitted laser pulse, ρ is the target reflectivity, A0 is the aperture area of the receiver, η 0 is the spectral transmission of the probe light, and γ is the atmospheric attenuation coefficient. 

According to the above formula, we can know that the peak power of light reception is directly proportional to the aperture area of the receiver. Therefore, the small mirror size of MEMS micro mirrors is a "hard flaw" that greatly limits the aperture of MEMS LiDAR when receiving signals, and the peak power of light reception is also difficult to meet the requirements!

Research institutions and enterprises have also proposed many experimental solutions to address the inherent problems of MEMS LiDAR. For example, in the selection of light sources, 1550nm fiber laser is chosen; In terms of photodetectors, an array receiver is selected. Learn from the strengths and weaknesses of MEMS micro mirrors to create a MEMS LiDAR that can be used in vehicles.

For example, in early 2019, Hesai Technology released the MEMS LiDAR PandarGT 3.0, which selected a 1550 nm fiber laser. The human eye safety threshold of 1550 nm laser is much higher than that of 905 nm laser. Therefore, within a safe range, the laser power of the 1550 nm fiber laser can be significantly increased, thereby increasing the peak power of the receiving end and making it more suitable for long-distance detection.

Physical image of PandarGT 3.0 launched by Hesai Technology

Toyota Central R&D Labs, a laboratory under Toyota, has released a LiDAR system that uses MEMS micro mirrors and SPAD (Single Photon Avalanche Diode) arrays to measure distances of 20 meters (Opt. Express 20, 11863 (2012)).
Toyota Central R&D Labs uses SPAD array receiving method to achieve MEMS LiDAR 

In short, the track of autonomous driving has been opened up, and various laser radar technology routes are competing to pursue on this track. Although we are optimistic about the strength of MEMS LiDAR, can MEMS micro mirrors pass the assessment smoothly in the face of strict requirements for zero defects in automotive chips? Will there be a turning point in the plot of solid-state LiDAR? Waiting for time to witness! Before that, it is necessary for us to comprehensively learn and understand various laser radar technology routes.