An overview of LIDAR and its use cases in Autonomous Cars

As 6 LIDAR companies have gone public, either via SPACs or have been public for a while, I thought I would dig deeper into LIDAR and answer a few questions as I formulate my thesis on LIDAR. There have been over 80 companies funded in the LIDAR space and over $3.5 Billion in investments that have been made over the last 7 years.

The public companies are: Velodyne Lidar (VLDR), Luminar (LAZR), Aeva (IPV / AEVA), Innoviz (CGRO / INNV), Ouster (CLA / OUST) and Microvision (MVIS). This write up is going to focus on LIDAR overall, while the next few days I will focus on each of the companies.

Velodyne is an early leader in both the market and revenues, while the others are expecting to ramp up production of products and revenue by 2025.

I have taken a Frequently Asked Questions approach to this write up, focusing on questions I would have as a beginner in this market.

What is LIDAR? (LiDAR)

LIDAR (or LiDAR) stands for Light Detection and Ranging, the use of laser pulses to build a 3D model of environment around objets.

Just like SONAR (which uses sound) and RADAR (which uses radio waves), LIDAR uses pulsed laser to measure distances and map objects.

Why is LIDAR suddenly important?

LIDAR has been around since 1950s and has been used by ecological organizations to track tree growth in forests and other sprawling ecosystems. Airplanes equipped with LIDAR soar overhead, and – even when travelling at cruising speed – can catalogue individual leaves and branches by the billions in a single sweep.

Filmmakers have incorporated this technology to chart rare environments or objects that cannot be used in production – like consecrated ruins and fragile artifacts – to recreate them in CGI

In the automotive industry, radar has long been utilized to automatically control speed, braking, and safety systems in response to sudden changes in traffic conditions.

The move towards autonomous cars (level 5) requires multiple sensors on cars – cameras, radar sensors, etc.

Whereas vision and RADAR-based sensors can provide a high level of automation to vehicles, the conception of a fully automated self-driving vehicle requires LIDAR-based sensors.

Self-driving cars equipped with LIDAR sensors offer complete automation under all driving modes. LIDAR-based ADAS systems along with vision and RADAR-based sensors take complete control over the vehicle, managing the speed and steering control, thereby providing a remarkably safe driving experience.

What is Level 5 autonomous driving?

The 6 levels of autonomous vehicles

Level 5 vehicles do not require human attention―the “dynamic driving task” is eliminated. Level 5 cars won’t even have steering wheels or acceleration/braking pedals. They will be free from geofencing, able to go anywhere and do anything that an experienced human driver can do. Fully autonomous cars are undergoing testing in several pockets of the world, but none are yet available to the general public. 

What types of sensors are needed for autonomous cars?

Automobiles need radars, cameras and ultra sound sensors for different purposes. Some for assisted parking, staying in lane, “sensing other cars ahead”, and viewing potential obstacles.

How does LIDAR work?

  1. Lidar systems emit a constant stream of laser pulses in every direction.

2. Laser pulses travel outward until contacting an object.

3. Upon contact the pulses reflect to the LIDAR system.

4. Lidar system received and registers the echo pulse.

5. The system calculates the distances of the object, based on the laser pulse’s travel time to and from it.

Lidar systems emit and receive hundreds of thousands of laser pulses every second.

A LIDAR system consists of four key components: a transmitter to emit laser pulses, a receiver to intercept pulse echoes, an optical analyzing system to process input data, and a powerful computer to visualize a live, three-dimensional image of the system’s surroundings.

How did LIDAR become important for autonomous cars?

LIDAR allows you to generate huge 3D maps, which you can then navigate the car or robot predictably within. By using a LIDAR to map and navigate an environment, you can know ahead of time the bounds of a lane, or that there is a stop sign or traffic light 500m ahead. This kind of predictability is exactly what a technology like self-driving cars requires and has been a big reason for the progress over the last 5 years.

How big is the market?

The LIDAR industry is projected to be worth $1.8 billion in three years and over $8B by 2025. The market for LIDAR sensors in light-duty vehicles could reach $46 billion in sales by 2030, with much of that going to enable partial autonomy.

LiDAR industry: high expectations for autonomous driving - System Plus  Consulting

Is LIDAR really needed? Can you do the same with cameras and software?

Critics of LIDAR including Tesla CEO Elon Musk believe the high cost of LIDAR (early units were > $75K), makes then untenable for autonomous cars. Musk has said that AI + cameras would suffice for autonomous cars.

LIDAR has the potential to make autonomous systems safer, If there’s bright sunlight or if it’s really dark at night, camera-based features are not reliable and not available all the time.

Advanced driver assistance systems (ADAS) are becoming more commonplace in mass-market vehicles. These systems rely on cameras and radar to power features like automatic emergency braking, blind spot detection, and lane-keep assistance.

For inclusive vehicle safety solutions, ADAS systems cannot be dependent on just vision and RADAR based systems; they require more efficient systems capable of providing highly accurate data for improved driver assistance.

Whereas vision and RADAR-based sensors can provide a high level of automation to vehicles, the conception of a fully automated self-driving vehicle is impossible without LIDAR-based sensors.

Self-driving cars equipped with LIDAR sensors offer complete automation under all driving modes. LIDAR-based ADAS systems along with vision and RADAR-based sensors take complete control over the vehicle, managing the speed and steering control, thereby providing a remarkably safe driving experience.

What are the types of LIDARs?

1. Electro-Mechanical LIDAR

Electromechanical LIDARs are traditional LIDAR systems, which can be considered as first generation LIDAR sensors for automotive applications. These mechanical spinning LIDAR system sensors are assembled from multiple moving parts, which are arranged to produce and emit an array of laser beams towards the targeted area. Electromechanical LIDARs are quite bulky, very expensive and prone to wear and tear in tough terrain. They are installed on the top of the vehicle and continuously rotate to scan the surroundings of the vehicle and typically cover a long range.

How does Google's Waymo self-driving car work: GRAPHIC - Business Insider
The Waymo Google Lidar is a bulky electro mechanical system

2. Solid State LIDAR

Unlike traditional electro-mechanical LIDARs, solid state LIDARs are built entirely on a single chip (e.g. AEVA or Luminar). All the components of the LIDAR systems like emitter, receiver and processors are integrated on the single chip of the solid-state LIDAR. Being on microchips, solid state LIDARs are compact in size. Also, they are not visible upon installation, light in weight and cost efficient. Since there are no moving parts in these LIDAR systems, they are fixed in rear, front and sides of the vehicle.

Solid state LIDARs have optical emitters, which send a burst of laser photons without having to adjust the direction of the transmitters. The light, emitted in specific patterns, collides with the objects in the way and bounces back to the system’s receiver. The processor in the LIDAR system fetches this data constantly and produces a real time 3D map of the vehicle’s surroundings.

Solid state LIDAR is the future of LIDAR based ADAS systems as it is exceptionally durable, reliable, affordable and commercially viable.

LeddarTech Releases 3D Solid State LiDAR System-on-Chips to Automotive  Partners - News
Cross Sectional view of LeddarTech solid state LIDAR

What are the alternatives?

Tesla is pushing the idea of pseudo-LIDAR based on innovations in cameras, AI and computing, and the fact that they have access to rich driving data from ~1M cars in varied traffic and weather conditions. 

Daimler has research activity, aimed at exploiting gated imaging and machine learning to replace LiDAR. GM Cruise stated in a recent interview that they had replaced one of the sensors in the sensor pod of their newly launched purpose-built ride hailing AV, the Origin, most likely implying that LiDAR will be replaced by thermal cameras.

Other methods to extract 3D information using stereo vision cameras have also been proposed.

What is the future of LIDAR?

The industry is marching ahead with a real focus on cost decrease and resolution and range increase.

Solid-state LIDAR opens the potential of sub-$1k powerful LIDAR units, which today can cost as much as $80k a unit. The huge jump in the number of applications for LIDAR has brought with it a flood of talented founders and teams starting companies in the space. Higher resolution output and increased tracking range (200m in some cases) will provide better object recognition and tracking and are one of the key differentiators in sensors from startups like Luminar.

References

https://www.theverge.com/2020/1/7/21055011/lidar-sensor-self-driving-mainstream-mass-market-velodyne-ces-2020

https://www.automotiveworld.com/articles/lidars-for-self-driving-vehicles-a-technological-arms-race/

https://www.forbes.com/sites/sabbirrangwala/2020/04/30/do-autonomous-vehicles-need-lidar/

https://www.forbes.com/sites/sabbirrangwala/2020/04/12/there-must-be-860-ways-to-build-an-av-lidarpart-1/#56ad2f25545c

https://www.ff.com/us/futuresight/what-is-lidar/

https://arxiv.org/abs/2004.08467

https://news.voyage.auto/an-introduction-to-lidar-the-key-self-driving-car-sensor-a7e405590cff

https://www.osa-opn.org/home/articles/volume_29/january_2018/features/lidar_for_self-driving_cars/

https://www.einfochips.com/blog/how-lidar-based-adas-work-for-autonomous-vehicles/

https://www.mdpi.com/2076-3417/9/19/4093/htm

https://www.ti.com/lit/wp/slyy150a/slyy150a.pdf?ts=1609138611013

https://www.ttnews.com/articles/fully-autonomous-driving-distant-lidar-finds-its-place

https://www.wired.com/story/lidar-self-driving-cars-luminar-video/

One thought on “An overview of LIDAR and its use cases in Autonomous Cars”

Comments are closed.