Any application for autonomous technology relies on an array of sensor modalities to perceive the environment around it. But no other modality plays as pivotal a role in artificial perception than light detection and ranging (LiDAR). Exceeding the performance of the human visual cortex requires a reliable, high-performance sensor such as LiDAR. MEMS (micro-electromechanical system) LiDAR has gained acceptance as the preferred approach due to its superior range, rate, and resolution performance capabilities. However, not all MEMS are created equal.
AEye’s innovative and patented approach to LiDAR makes our MEMS the most robust and ruggedized available for autonomous cars, trains, trucks, construction equipment — or any application that experiences a lifetime of shock and vibration. In testing completed at NTS, one of the most respected testing, inspection, and certification companies in the US, AEye’s sensor, 4Sight M, surpassed automotive standards for both shock and vibration. See related press release.
So how does AEye create the most robust MEMS on the market? Here are a few secrets to how we did it:
The device has to have a high natural resonance frequency
This all depends on what kind of excitation the environment is under. For the automotive vibrations spectrum, we have a large amplitude of vibrations in the 10s to 100Hz range, but that goes down an order of magnitude when you slowly reach kilohertz.
We must also consider the differences between a sprung system and an unsprung system. For sprung systems, the amplitudes are much less. So, if you want to be robust against that, your device has to have a natural resonance frequency that’s higher so that it doesn’t get too excited by those frequencies.
Mirror size
Not only do larger mirrors have a lower resonance frequency (meaning that they experience more shock and vibration amplitude, which can be orders higher than smaller mirrors), larger mirrors also have larger inertia, generating 10x to 600x more torque from shock and vibration events. Simply put: the bigger the mirror, the easier for inertia (acceleration) to move in and swing it.
The combination of both these factors result in a mirror which can easily be deformed beyond yield strength for flexures, leading to failure during operation. Think of a bank vault door. Although it’s large and heavy, it’s actually quite easy to open due to torque force. In the same way, large mirrors are more susceptible to resonances: they’re easy to turn, therefore, they will swing more.
In addition, larger mirrors do not allow for fast, quasi-static movement for active scanning, which is key to intelligent and reliable artificial perception.
Flexure structure and stiffness
Generally, when you have large mirrors, you want them to be fast. This means they must be stiff. But if you have a stiff structure that swings and flexes a lot, it may ultimately go beyond its yield strength and break.
However, some mirrors are designed to not be very flexible (as long as those external excitations are not too large). They won’t necessarily break, but they will vibrate quite a bit.
AEye’s innovative approach to MEMS
The unique design of AEye’s MEMS has a mirror that is approximately 1mm in size, compared to the industry average of 3–10mm. While there are several important trade secrets in the AEye design, the smallness of the MEMS are one of the key ways AEye’s MEMS are optimized to deliver the unique high-performance of iDAR™ with ground breaking reliability. Because our mirrors are so small, we can achieve high frequency without stiffness, as smaller mirrors with higher resonance are more robust against shock and vibration. Extensive shock and vibration testing at NTS showed that AEye’s 4Sight sensor can sustain a mechanical shock of over 50G, random vibration of over 12Grms (5–2000Hz), and sustained vibration of over 3G for each axis.
Our competitors must have larger MEMS or MEMS arrays because of their traditional system design — they are monostatic (e.g. transmit/receive signal). However, because we set out to build an active, intelligent sensor, and we use a patented bistatic receiver concept with parallax correction, our MEMS can be extremely small, enabling us to be very robust against environmental factors.
Some competitors utilize multiple transmit paths to achieve more flexibility. But this adds cost, complexity and increases risk of failure because you have introduced many more devices that can fail. Multiple laser arrays are also fundamentally limited as they add many calibration challenges on top of cost and complexity.
The bottom line is this: while there are numerous approaches to creating MEMS-based LiDAR, when you have large or multiple mirror arrays, you must compensate in ways that compromise price, performance, and/or reliability.
Hope this has helped explain why all MEMS are not created equal
AEye is changing the calculus for adding high-performance LiDAR to a vehicle. Now, automotive OEMs and Tier 1s — along with trucking, construction, transit, rail, ITS, aerospace, and defense markets — can implement LiDAR with the confidence that it can withstand forces similar to those generated by the recent historic launch of a Falcon rocket. In addition, AEye’s use of standard micro-MEMS in a custom design and standard manufacturing processes allows the overall solution cost to scale like other solid-state systems.
AEye is also paving the way for accelerated innovation in autonomous technology. With mirrors robust (and small) enough to allow for active scanning and software -configurability — they are the key to the development of intelligent and reliable next generation artificial perception.
To experience AEye’s 4Sight sensor and receive a live demo, please email 4sight@aeye.ai.
About AEye
AEye is the premier provider of high-performance, AI-driven LiDAR systems for vehicle autonomy, advanced driver-assistance systems (ADAS), and robotic vision applications. AEye’s smart, software-configurable iDAR™ (Intelligent Detection and Ranging) platform combines solid-state, active LiDAR, an optionally fused low-light HD camera, and integrated deterministic artificial intelligence to capture more intelligent information with less data, enabling faster, more accurate, and more reliable perception. The company is backed by world-renowned investors including Kleiner Perkins Caufield & Byers, Taiwania Capital, GM Ventures, Intel Capital, Continental AG, Hella Ventures, LG Electronics, Aisin, Airbus Ventures, SK hynix, Subaru-SBI, and Tyche Partners.
Learn more at www.aeye.ai
