LiDARs based on 900 nm are not eye-safe and don’t support FMCW; 1550 nm supports FMCW but is too expensive; FMCW and 1310 nm are “a marriage made in heaven.”
LiDAR, which stands for “Light Detection and Ranging,” is a method for determining the distances of objects (and, possibly, additional information associated with those objects) by targeting them with a laser and measuring the time for the reflected light to return to the receiver.
In the case of LiDAR sensors used for automotive applications, the majority fall into two main wavelength bands. One group is found around 900 nanometers (nm), while another group is found at 1550 nm.
So, why did we at Insight LiDAR decide to create our state-of-the-art LiDAR systems using a wavelength of 1310 nm? In order to answer this question, let’s first consider what drove the creators of traditional automotive LiDAR systems to opt for 900 or 1550 nm lasers.
Power vs. Distance vs. Safety
In the case of what we might call the 9XX band (different versions may be found at 852 nm, 905 nm, 980 nm, etc., all using the same material and same type of laser tuned to different frequencies), engineers were already using these lasers for a wide variety of existing applications. So, when designers started to consider using LiDAR for automotive applications, one reason for selecting the 9XX wavelengths was that there were already so many of these lasers available.
Unfortunately, there are two main problems with the 9XX wavelength for automotive applications. The first is that the human eye conveys light at this wavelength directly to the retina, where it can cause significant damage. The last thing anyone wants to do is build a car that blinds drivers or any pedestrians it passes in the street, so this limits the amount of power that can be used at this wavelength while still being considered to be “eye safe.” However, limiting the power limits the range over which the LiDAR can “see.”
As an alternative to the 9XX range, many designers decided to use the 1550 nm wavelength, which is widely used in fiber optic communication systems and other photonics and optical components. Using 1550 nm allows you to transmit more power while still remaining eye-safe, thereby allowing these automotive LiDAR systems to see a lot further than their 9XX counterparts.
However, “there’s no such thing as a free lunch,” as the old saying goes. The difficulty here is that in order to produce the required amount of power, it’s necessary to add a fiber amplifier, which is relatively large and expensive and which itself consumes power.
All of this explains why we at Insight LiDAR decided to base our systems on a 1310 wavelength as an ideal compromise. At this wavelength, our LiDAR systems can put out a lot of eye-safe power that allows them to see farther. In addition, we can achieve this power by means of on-chip semiconductor amplifiers, which is a much, much cheaper solution than what can be achieved using fiber amplifiers, especially at scale.
TOF vs. FMCW
Another thing to consider when creating automotive LiDAR systems is that most of the systems in use today employ a detection technique called time-of-flight (TOF). This involves generating short, powerful pulses of light and looking for their reflections. In turn, this means you need a laser source that can produce short, powerful pulses.
Lasers that can generate high-powered pulses of this type exist in the 9XX and 1550 bands, but these really aren’t available for the 1310 nm wavelength. This means that 1310 nm is not a good choice if you wish to create a TOF LiDAR, so as a result those creating TOF LiDAR sensors fall into the 9XX or 1550 nm camps.
A more sophisticated detection technique is known as frequency-modulated continuous wave (FMCW). As well as requiring much less power than its TOF counterpart, this also provides more information. In addition to distance, by means of the Doppler effect, it’s possible to determine the approaching or receding velocities of objects on a pixel-by-pixel basis. As fate would have it, this capability is not supported by 9XX nm lasers. Also, although FMCW is supported by 1550 nm lasers — as we’ve already discussed – these are too expensive for widespread commercial deployment and therefore provide no pathway for affordable deployment at the volumes required by the automotive industry. However, FMCW and 1310 nm lasers are “a marriage made in heaven” by solving both the eye-safe challenge at the same time you’re putting the affordability question to rest.
Reflection and Absorption
Another interesting aspect with respect to creating FMCW LiDARs for automotive applications is that most of the objects we are interested in seeing — clothing, people, animals, trees – are more reflective when using the 1310 nm waveband than they are at the 1550 nm waveband.
This difference can be significant, with many objects being 20% to 30% more reflective (in some cases we are talking 4X or 5X more reflective) at 1310 nm than at 1550 nm. For example, 1550 nm signals are absorbed by water. As a result, human skin appears dark when illuminated at 1550 nm, but remains reasonably reflective at 1310 nm.
Another aspect of all this is environmental conditions. Once again, since the 1550 nm wavelength is strongly absorbed by water, its useful detection range is significantly degraded by fog or snow. By comparison, LiDAR systems using the 1310 nm wavelength are comparatively unaffected. In fact, 1310 nm LiDARs can “see” better than their human counterparts.
The Road Ahead
At Insight LiDAR, our current 1310 nm systems can scan a full 120- by 30-degree field of view at 10 Hertz, providing a scan 1,000+ points wide by 300+ points deep. Our technique produces many, many more points than competitive solutions. More points in turn means higher resolution, which means you get to see things better and farther than when using alternative approaches. In fact, in addition to providing the distance and instantaneous velocity associated with every pixel, and in addition to fielding an all-semiconductor architecture, Insight LiDAR’s technology boasts 10-100x better sensitivity, 64x higher resolution, and 10,000x better rejection of sunlight, LiDAR, and other light sources. It’s the clear engineering solution for the scalable, reliable ADAS future ahead.