Wednesday, July 22, 2020

Value Question: Lidar

Thanks Snowboard for the assist! 

As always, I show my work... (and I believe the exercise involved in finding a method is more important than the result.) This is all my own work, this is an exercise, Im not giving any investing advice, do you own homework, (etc.)

Coming up with a concrete number here is very difficult. There is little if any price/value history on LiDAR, and much of what is there is from experimental usage.

This is probably the most valuable part of the company -- The CEO is there because of LiDAR.


Coming up with any numbers here is difficult. Lots of guesswork  (I'm admitting it, the best we can get is a SWAG.)


A few assorted facts:
  • All LiDAR will be using class 1 lasers. (Even though you can't see infrared laser light, stronger than class one could still damage eyes, and there is no blink reflex for infrared light.)
  • MicroVision's target is 20,000,000 points per second, that's a lot of points. You can count hairs on people with that.
  • The significant limiting factor on range is the Speed of light. (see chart below)

A few main areas of likely use: (There are more, like SLAM, scanning and 3D printing... )
  • Gesture Control
  • Security
  • Adas
Gesture control:
This could be huge....
...but after wrestling with this off and on for days, I really can't come up with a defensible guess on it.

The potential is enormous though. You could, for example build a new home that has no wall switches at all, and have anything you want controlled by gesture. (Turn on the lights, raise or lower the blinds, adjust the temperature, etc.) With one sensor and one wire instead of multiple switches and needing to get up to do it.

I could come up with something, but the SWAG would emphasize the wild-ass part of it. After already exerting considerable effort in this direction, I've determined that further effort will not result in a good and measurable estimate of a future reality.


No Guess 

Security

Subscriptions + Hardware.

I've seen MVIS's LiDAR in action at a shareholder meeting. It can recognize people. This has been described on MicroVision's conference calls, and has been described with significant additional safety and convenience features.

  • This could identify individual people
  • Can distinguish between pets and people (or YOUR pet and the neighbors pet)
  • Can distinguish between normal behaviors and strange things that could be of concern
  • Could save face-scans of intruders and allow intruders to be identified later
Home Security Subscribers -- Approximately 10 Million last year

Would a MicroVision LiDAR improve a Home Security system setup? Would it improve the services one of those companies can provide?  

I'm a definitive yes on this.


I'll assume that 25% would like and pay for the capabilities, and they could perhaps add 25% of subscriptions. (Identifying people in the home would be huge, along with other benefits.) At an approximate $75/month additional to another subscription with 1/3 of that credited directly to the ability of the MicroVision tech. (Still a SWAG, remember.)


If yes, 25 * 5,000,000 = 125Million per month. = 1.5 Billion per year. Over a 5 year window you could run that out to 7.5 Billion ($53/share)

Like Beats, selling the subscription is the money center.


ADAS

What is the value of preventing accidents? It's huge in the amount saved in damage and life.

Will your car insurance rate change with ADAS

If you're using a self-driving car and it crashes, who's fault will that be? -- the company that programmed it and designed the system. (They'll obviously put a self-testing system in place -- and it won't move unless their system is working... or you'll have to disconnect it and take responsibility yourself.)

Car insurance premiums could drop by as much as $25 billion, or 12.5% of the total market, by 2035 thanks to the rollout of autonomous vehicles, consulting firm Accenture noted in its report.

Accenture calls it an $81 Billion opportunity. (If you give MVIS a 5% share of that number is $4 Billion or $28/share)

Government requirement. (Something small, cheap easy to install, high resolution is a GOOD thing to have in this market.)

What is the value of a self driving vehicle.
What is the value of a vehicle that is mostly ding-and dent proof?

American Family Insurance: Discounts for:

Forward Collision Warning

This car safety feature tracks a vehicle’s speed and compares it to the vehicle ahead. If the distance is closing too quickly, the system warns the driver to slow down through a visual alert, audible warning, or combination of the two. This warning gives drivers time to brake and avoid a collision.

Automatic Emergency Braking

If the driver doesn’t respond to the forward collision warning, some vehicles also have automatic emergency braking. This car safety feature automatically applies the brakes to slow or even stop the vehicle. Depending on your speed, automatic emergency braking can reduce the severity of a crash or even prevent it completely.

(AT Consumer Reports)


Lane Departure Warning and Lane Keep Assist

The lane departure warning system uses cameras to see road markings. If a driver drifts out of the lane without signaling, then this feature issues an audible or visual warning so the driver can steer the car back into its lane. Some systems also vibrate the driver’s seat or steering wheel to get the driver’s attention. Working along with lane departure warning on some cars is lane keep assist. This will gently nudge the steering wheel in the correct direction to help the driver stay in his lane.

Blind Spot Detection and Rear Cross Traffic Alert

When there’s a vehicle in the blind spot, this car safety feature alerts the driver with an audible or visual warning. Some systems issue a warning only if a car is in your blind spot when you use your turn signal while other systems activate a visual alert any time a car is in the blind spot, even if you’re not signaling to turn. Rear cross traffic alert is another system to help you see what you might miss. When the car is in reverse, it monitors traffic about to drive behind the vehicle and warns the driver of any oncoming cars.

SLAM Simultaneous Localization and Mapping

When robot's start moving among us, this will be very important.

This could have enormous impact in tool and die, manufacturing, etc, but it's very difficult to put a number to it. (Just making sure your tool doesn't crash into something it shouldn't this could be big.

Could heavy equipment scan the area for fragile items and LiDAR in concert with onboard software ensure that this equipment doesn't move anything it shouldn't? Probably yes, probably would save a lot. (Could a crane operator, for example move loads MUCH faster with much more confidence?)

No history --- no guess... but someone out there has a value for this..

Scanning / 3D printing

No guess.... this is another one of those items that could be a very big market, or very small... no way to know yet.

No guess. (There is value for this too.)

The RANGE AND RESOLUTION CHART


Included because it's interesting -- and because it provides understandings of important limitations.

The laser in a lidar must be class 1 (brighter could still damage eyes, even though the light can't be seen)

Range and resolution are limited by the speed of light.

I produced the spreadsheet below after listening to a bunch of engineers at a STM conference argue about LiDAR range. They spent some time arguing that the speed of light was a limiting factor. Since it's the fastest thing, I did this to understand that argument.

For each pixel of resolution, the light must leave the device, bounce off the object at the end, and return to the source.


Constants
All of this is a HARD LIMIT, not of the hardware, but a hard limit based on the speed of light.
ConstantValueScaleName
Light Second299,792,458meters/secLightSec
720P921,600PixelsLowHD

Lightsec div PS
1080P2,073,600PixelsHD
Points / SecondRange in METERS
Refresh60HertzRefreshHD X Hertz124,416,0002.41

Adjusting the Refresh rate (scans per second) changes the effective range (Using all the pixels of that standard resolution.) For 60 Hertz, we could use a number less than 60hz to account for error-rates, etc.
Low HD X Hertz55,296,0005.42

for points per second it only requires the number of points and the speed of light to find the distance.
20,000,00014.99
10,000,00029.98
5,000,00059.96
1,000,000299.79
These range numbers assume 60hz






1 comment:

  1. Bravo! You should get an honorary degree from Wharton, my alma mater.
    Over the years you have been a breadth of fresh honor with good research
    and opinion. Thank You.

    ReplyDelete