Mobileye Expands Autonomous Vehicle Testing to New York City | Intel
Mobileye Expands Autonomous Vehicle Testing to New York City
Intel’s subsidiary, Mobileye, has broadened the scope of its autonomous vehicle testing program with the inclusion of New York City. This expansion is a key component of the company’s overall plan for the development and eventual deployment of self-driving technology.
Global Testing Network
New York City now joins a growing list of global locations where Mobileye is actively testing or intends to test its autonomous systems this year. These cities include Detroit, Paris, Shanghai, and Tokyo. Initial testing began in Jerusalem in 2018, followed by Munich in 2020.
Challenging Environments are Crucial
“To create a system capable of widespread implementation, we must demonstrate its functionality in demanding environments, and essentially everywhere,” stated Amnon Shashua, Mobileye’s president and CEO, during a live-streamed presentation on Tuesday. Mobileye simultaneously released a 40-minute, unedited video showcasing a test vehicle navigating the complex streets of New York City.
Camera-Based Autonomous Driving
The vehicles currently undergoing testing in New York City operate autonomously, though a safety operator remains present behind the wheel. These vehicles rely exclusively on camera technology for their autonomous capabilities. They are equipped with eight long-range cameras and four parking cameras, all powered by Mobileye’s fifth-generation EyeQ5 system on chip.
Redundancy Through Sensor Fusion
However, Mobileye isn’t solely focused on a camera-only approach for full autonomy. The company has also developed a separate subsystem utilizing lidar and radar, without cameras, that also achieves autonomous driving. These two distinct sensor and software systems will ultimately be combined to provide redundancy for robotaxi services.
Consumer and Robotaxi Applications
The camera-only subsystem, as described by Shashua, is designed to be “at the cost level for consumers” and will be used to enhance existing driving assist systems. Later this year, Mobileye’s camera-only system, leveraging the EyeQ5 SoC, will be integrated into a vehicle from the Geely Auto Group.
Regulatory Approval and Unique Position
New York City has been a target for Shashua for over six months. He first expressed interest in public road testing in New York during the 2021 CES tech trade show, contingent upon receiving the necessary regulatory approvals. Now, having secured this approval, Mobileye is currently the only company authorized to test AVs within the state and city. GM’s Cruise previously outlined plans to test in New York in 2017, even mapping portions of lower Manhattan, but ultimately prioritized San Francisco for commercial deployment.
Compliance with State Regulations
Mobileye obtained a permit through New York State’s autonomous vehicle technology demonstration and testing program. The company fulfilled all program requirements, including adherence to federal standards, New York State inspection standards, and a comprehensive law enforcement interaction plan.
Testing Challenges in New York
“The approval process itself isn’t particularly noteworthy; it simply requires completing the necessary steps,” Shashua explained, characterizing the process as lengthy and comparable to the rigorous requirements for testing in Germany. “The true challenge lies in the difficulty of driving in New York City.”
Mobileye’s Core Business
Mobileye is widely recognized as a leading supplier of computer vision technology to automakers, powering advanced driver assistance systems. This core business generated nearly $967 million in sales for the company. Currently, 88 million vehicles on the road utilize Mobileye’s computer vision technology.
Full Self-Driving Stack
Alongside its ADAS business, Mobileye is actively developing full self-driving technology. This includes a redundant sensing stack based on camera, radar, and lidar, combined with its REM mapping system and a rules-based Responsibility-Sensitive Safety (RSS) driving policy.
Crowdsourced Mapping with REM
Mobileye’s REM mapping system leverages data crowdsourced from consumer and fleet vehicles equipped with its EyeQ4 chip to create high-definition maps. This data, which is compressed text rather than video or images, collects approximately 10 kilobits per kilometer. Mobileye has data-sharing agreements with six OEMs, including BMW, Nissan, and Volkswagen, and also collects data from an aftermarket product sold to commercial operators.
Global Mapping Coverage
Mobileye’s technology is currently mapping nearly 8 million kilometers daily worldwide, including within New York City.
Dual Strategy: Robotaxis and Consumer Vehicles
Shashua believes this strategy will enable the company to efficiently launch and operate commercial robotaxi services while simultaneously bringing the technology to consumer passenger vehicles by 2025. He detailed this dual approach in a 2020 interview with TechCrunch.
“We came to realize some time ago,” he stated, “that the ultimate goal is passenger car autonomy – the ability to purchase a vehicle with an optional autonomous feature that can transport you to any destination. The realization was that achieving this requires first establishing a robotaxi business.”
Unique Position in the Industry
On Tuesday, Shashua emphasized that Mobileye is uniquely positioned as the only company actively pursuing both robotaxi and consumer vehicle autonomy. (Though Toyota’s Woven Planet does exhibit some overlapping strategic goals.)
“We are developing our technology to support scalability, particularly geographic scalability, through our crowdsourced mapping technology and by creating new sensors. This ensures the entire system – the complete package – will be under $5,000, enabling consumer AVs. Simultaneously, we have a division focused on building a mobility-as-a-service, or robotaxi, service,” Shashua explained. “This is why we acquired Moovit last year, to manage the customer-facing layers above the self-driving system and facilitate the mobility-as-a-service business.”