Most people will recognize a self-driving car based on the whirling sensor perched on the roof. LIDAR, which stands for Light Detection and Ranging, uses laser pulses to build a 3D model of the environment around the car. Essentially, they help autonomous vehicles “see” other objects, like cars, pedestrians, and cyclists. Chris Urmson, the former head of Google’s self-driving car project, once famously described LIDAR as “spinning Kentucky Fried Chicken buckets.”
They are very cool, but also very expensive. A popular early model from Velodyne, used by many self-driving companies, cost $75,000. Hardly practical for mass production. Thankfully, they’re getting cheaper. So cheap, in fact, that the companies leading the pack now predict LIDAR will become as commonplace in mass-market vehicles as cameras, radar, and other low-cost safety technology.
“There is absolutely a movement afoot to add LIDAR to mainstream automotive vehicles, which is something that you would buy off of a dealership,” said Anand Gopalan, who earlier today was named as the new CEO of Velodyne. “They will start with the luxury vehicle brands, and then move their way downwards. We want to accelerate that process.”
There are still plenty of LIDAR that can break the bank, but the price for less powerful versions have gone down considerably. At the Consumer Electronics Show this week, Velodyne is showing off a new model called the VelaBit, which has a range of 100 meters and will cost only $100. It’s certainly not as powerful as some of the company’s top-of-the-line LIDAR, which boast ranges of up to 200 meters.
The VelaBit is designed to be used for autonomous vehicles, as well as manually driven cars with advanced driver assistance systems. It has a 60-degree horizontal field of view (FoV) and 10-degree vertical FoV. Its compact size — it’s smaller than a deck of playing cards — allows automakers to embed it seamlessly into their vehicles.
Velodyne isn’t the only company announcing a new sensor at CES. The show is lousy with LIDAR this year. Though technically not a consumer product, the presence of so many LIDAR products at CES speaks to the fervent belief that the sensor is poised to go mainstream. The LIDAR industry is projected to be worth $1.8 billion in five years.
Earlier this week, German auto supplier Bosch announced it was jumping into the LIDAR market with a new, long-range LIDAR with a wide field of view. This is significant, as Tim Lee notes at Ars Technica, because Bosch is a Tier 1 supplier, giving it the scale and resources to deliver LIDAR to the entire global auto industry.
Luminar, another LIDAR startup which has raised over $250 million to date, unveiled its own updated sensor, as well as a subscription model. The company says its new Hydra LIDAR can “see” objects that are as far as 250 meters away and boasts an overall range of 500 meters. And DJI, a Chinese company known for its consumer drones, said it was investing in a LIDAR company called Livox that claims a superior method of scanning objects.
All the LIDAR companies are trumpeting the same message: better resolution, wider FoV, longer range, more precision, and cheaper, cheaper, cheaper. If LIDAR is supposedly a crucial component in how self-driving cars perceive the world, then the LIDAR companies want the traditional automakers to see this potential, too.
The chances are high that they will succeed in this quest. Advanced driver assistance systems (ADAS) are becoming more commonplace in mass-market vehicles. Most of these systems rely on cameras and radar to power features like automatic emergency braking, blind spot detection, and lane-keep assistance. Sometimes they work really well, helping prevent a wide variety of crashes and taking some of the tedium out of driving. Other times they can be dangerous and even deadly.
The American Automobile Association recently conducted a series of tests using vehicles with automatic emergency braking and pedestrian detection alerts on a closed course with dummy pedestrians. Their findings were alarming: the dummies were struck while crossing the road 60 percent of the time, during daylight hours at speeds of 20 mph.
LIDAR has the potential to make these systems safer, said Velodyne’s Gopalan. “If there’s bright sunlight or if it’s really dark at night, some of these [camera-based] features are not reliable and not available all the time,” he said. “Adding a LIDAR will make two features much more reliable and much more readily available.”
Not everyone believes that LIDAR is a surefire solution. Tesla CEO Elon Musk has made his disdain for the sensor widely known. LIDAR is a “fool’s errand,” he said at an event for investors last year, and “anyone relying on LIDAR is doomed. Doomed. Expensive sensors that are unnecessary. It’s like having a whole bunch of expensive appendices.” Musk has argued that a camera and radar-based system, coupled with a powerful AI software, can compensate for the lack of laser sensing.
Gopalan thinks Musk is relying on “out of date” information about LIDAR. It’s true that a decade ago LIDAR sensors were too big and too expensive to make them viable for mass consumer application. But things have changed, he argues. “Frankly that ship has sailed,” Gopalan said. “And I think Elon is out of touch.”