LiDAR What is LiDAR, and why do self-driving cars need it?
As cars begin to drive on their own, obstacle detection and avoidance remain critical technical hurdles standing in the way of fully autonomous operation. LiDAR is one technology that can enable this, but the development of low-cost LiDAR systems is still in its infancy.
Similar to radar, LiDAR is a measurement and range finding technique originally developed in the 1960s. Useful for military and aerospace applications where precise accuracy over long distances was needed, this technology ultimately found its way into meteorology and robotics.
But LiDAR's potential application in vehicles portends a much larger marketplace that could be worth as much as $2.5 billion on its own by 2030.
This article is a brief overview of how LiDAR works and its applications in vehicles.
Where does the word LiDAR originate?
LiDAR stands for Light Detection and Ranging; originally, the word was a portmanteau of the words "light" and "radar." LiDAR has many applications, but it's particularly applicable to self-driving cars. In fact, it's likely that without LiDAR's capabilities, self-driving vehicles could not progress much further than what the Society of Automotive Engineers (SAE) terms Autonomy Level 3 capabilities, or limited "eyes-off driving."
How it works
LiDAR is a combination of laser scanning and 3D imaging. Car-based LiDAR systems make use of invisible, eye-safe laser beams (low-power ones, like those used at grocery store checkout counters). These are emitted from transmitters attached to a vehicle and bounce off all objects located within a maximum of 500 meters. Such objects can be either stationary or moving. A typical system, such as those supplied by industry vendor Velodyne, uses between eight and 128 individual laser beams to perform this function.
Processors on the vehicle time how long the beams take to reflect off the objects and return to the LiDAR's sensors. Then, the system uses formulas to calculate how far the objects are away from the vehicle.
By emitting many beams and constantly scanning the surrounding environment in 360 degrees, one or more LiDAR sensors can compile a 3D image of everything surrounding the vehicle they're attached to.
By getting distance data for every surrounding object, a LiDAR system can detect if a collision is imminent and warn a vehicle's driver as early as possible. In the near future, signals could be sent to autonomous driving systems to automatically activate steering, braking, acceleration, and other mechanisms. Thus, a vehicle could intelligently navigate around obstacles and/or forestall collisions. Even if objects are moving (pedestrians or other vehicles, for instance), their speed and trajectories can be calculated and taken into account.
Why LiDAR is superior to other technologies
Because of the way it functions, LiDAR is often seen as superior to other environmental sensor systems that rely on video cameras or radar. Both of these latter methods are used for autonomous Advanced Driver-Assistance Systems (ADAS) functions, including adaptive cruise control, lane-keeping, and parking.
One reason for this superiority is because both cameras and radar still have blind spots. Another reason LiDAR can be better is that it scans at extremely high resolution, and with decent computing intelligence, it can potentially identify numerous short-range objects.
This identification can filter disparate objects into those that might pose a threat (such as anything moving, for instance) and those that are simply part of the surrounding landscape.
A third reason for LiDAR's superiority is that, depending on how advanced it is, it can work at night or under harsh weather and environmental conditions, such as snow, rain, hail, fog, smoke, etc.
Essentially, the combination of LiDAR sensing and intelligent decision-making approximates how the human brain processes locomotion. Regardless of whether a person is a driver in a vehicle, a rider on a bicycle, or simply a pedestrian on foot, the brain is constantly taking in visual information and cues and reassessing risks from external objects and obstacles. It then tells the body to take proactive or reactive action as necessary to maintain safety while it continues to move forward.
Most people probably take for granted how much sensory information the brain processes and how many separate decisions are involved in daily navigation, but it's a lot in each case.
Examples of where LiDAR is beneficial
The faster a vehicle is moving—and the faster other moving objects are approaching—the more valuable a computer-assisted system like LiDAR becomes. An extreme example of this is a highway interchange or four-way intersection where other vehicles are approaching, often at high speed and from multiple directions.
In situations like this, the ability to differentiate between, say, a motorcycle with a rider and a large piece of garbage blowing across the road, is critical. LiDAR's high resolution combined with the intelligence of a microprocessor can not only protect the life of a vehicle's driver, but it can also safeguard the lives of those persons traveling or navigating nearby.
Another instance where LiDAR could be beneficial would be changing lanes or passing other vehicles on a road. Although these processes seem simple, the amount of "processing" the brain does to be able to affect these actions is substantial. Judging the distance and speed of other moving vehicles is critical to timing steering, braking, and acceleration correctly.
In short, there's no doubt that LiDAR will play a crucial role in autonomous vehicles of the future. As this technology matures, it will continue to get smaller, cheaper, and ever more powerful. Some industry players are now developing solid-state single or array "on-a-chip" LiDAR emitters that feature few or no moving parts, whereas current systems rely on rotating pieces to move their units' lasers around. It's likely that this newer solid-state form of LiDAR will supplant the rotating kind and continue to drive costs down.
The second article in this series explores the major players, costs, and business shakeups of the LiDAR industry over the last five years.