Mobileye’s CEO issued an unusually frank appraisal of the state of autonomous driving development, arguing that the cost of a self-driving system (SDS) for a passenger car needs to be one-tenth what it is today.
Mobileye CEO Prof. Amnon Shashua also said Mobileye endorses a system of deploying cameras for helping guide autonomous vehicles (AVs) rather than “over-engineering” with a range of sensors, such as lidars and radars for guidance.
On top of costs and other concerns, regulatory matters remain the “stickiest issue” because of the “serious issue of how to balance safety and usefulness in a manner that is acceptable to society,” said Shashua in a blog on July 9.
Shashua said the cost of an SDS is currently in the tens of thousands of dollars and will remain so for the foreseeable future. That covers the cameras, radars, lidars and high-performance computing needed for a single vehicle. “This cost is acceptable for a driverless ride-hailing service, but it is simply too expensive for series-production passenger cars. “The cost of SDS should be no more than a few thousand dollars.”
Under the heading of regulation, he argued that it will be easier to pass laws governing fleets of robotaxis than for privately-owned vehicles. “Licensing [SDS] cars to private citizens will require a complete overhaul of the complex laws and regulations that currently govern vehicles and drivers,” Shashua added.
“The auto industry is gradually realizing that autonomy must wait until regulation and technology reach equilibrium and the best place to get this done is through the robotaxi phase,” he said.
While some technology analysts and government officials have questioned the tlmeline for rolling out SDS cars in the next five to ten years, very few have wanted to kibosh the enthusiasm of the car and tech industry for SDS. Major companies and some states have invested millions in long-range testing programs and feel the need to defend their efforts. SDS “is a lot of hype so far,” one auto industry analyst recently said in an interview, asking to remain anonymous.
Shashua’s remarks were measured, although they were primarily intended to make a case for using SDS in the robotaxi industry. “When the factors of cost, regulation and scale are taken together, it is understandable why series production passenger cars will not become possible until the robotaxi phase,” Shashua said.
“At Intel and Mobileye we are all-in on the global robotaxi opportunity,” he said. The companies are developing technology for the robotaxi experience, including hailing a ride on a smartphone and powering vehicles and monitoring fleets.
Rather than pushing for SDS in passenger cars as has happened in the past, Shashua said the auto industry is “gravitating” towards Enhanced ADAS (Advanced Driver Assistance Systems), where drivers are still left in charge of the vehicle at all times. That level of automation helps achieve safety benefits of autonomous vehicles without facing the regulatory and cost challenges. Automakers, instead, are embracing the emerging robotaxi mobility as a service industry.
He also laid out the Mobileye strategy for autonomy in four areas:
- Continue ADAS development, now reaching 34 million vehicles on roads today.
- Design SDS solely around a camera-centric configuration “to avoid unnecessary over-engineering or sensor overload” from radars and lidars.
- Build on Mobileye’s Road Experience Management crowdsourced and automatic high-definition map-making. Mobileye expects to have more than 25 million cars sending road data by 2022.
- Tackle regulatory issues with Mobileye’s Responsibility-Sensitive Safety (RSS) model of safe driving. RSS is described by Mobileye as a “formal mathematical model to help ensure that an autonomous vehicle is operated in a safe manner.”
RSS offers specific and measurable parameters designed to mimic human concepts of caution in driving. It defines a “safe state” to prevent an AV from causing an accident “no matter what action is taken by other vehicles.”
In a video animation from Intel, an example of how RSS would apply to traffic situations shows an AV refusing to merge onto a busy freeway until a specified, safe gap between cars is sensed.
RELATED: "A robust market for LiDAR sensors"