Waymo Releases Report on How it Ensures its Autonomous Vehicles Offer the Highest Level of Safety

author: FutureCar Staff    

A self-driving Waymo minivan.

As the company that spun out of Google's self-driving car project, autonomous driving developer Waymo has been working on self-driving cars for over a decade and is considered the industry leader. Since 2009, the company has logged over 20 millions miles with its fleet of self-driving vehicles and billions more in computer simulation to ensure its self-driving system offers the highest level of safety.

As a leader in the development of autonomous driving technology, including commercial robotaxis and self-driving trucks, Waymo developed comprehensive safety requirements for its autonomous vehicle development, testing and deployment phases which the company shared in two detailed reports.

Waymo's first report is titled "Safety Methodologies and Safety Readiness Determinations" and includes details about the company's multilayered approach to safety that will make it possible for Waymo to deploy fully-autonomous vehicles on public roads at scale. The other is titled "Waymo Safety Report."

Currently, Waymo's autonomous vehicles are mostly deployed in California and Arizona, but the company performs additional testing in several other states including Michigan,Texas, Florida and Washington. In Arizona and California, Waymo is testing vehicles that will eventually be used in a commercial ride-hailing service called Waymo One.

Waymo's safety methodologies focus on the development, qualification, deployment and field operation of a Level 4 autonomous driving system (ADS) designed to operate without human intervention in both normal traffic conditions and very challenging urban environments. The company's methodologies address the full life cycle of its self-driving vehicles from design and development, through phases of testing, and during and after deployment.

Waymo's early testing as Google's self-driving car project relied on human intervention for an added layer of safety, but with its goal of Level-4 autonomy that requires no human intervention, Waymo is now enhancing safety through automation.

Waymo refers to its entire ADS as the "Waymo Driver."

In 2017, the U.S. Department of Transportation (DOT) outlined twelve safety design elements and encouraged companies testing and deploying self-driving systems to address each of these areas in a report called "Automated Driving Systems 2.0: A Vision for Safety."  

Waymo addressed each of these and outlined the processes for each safety design element and how they support the development, testing, and deployment of fully self-driving vehicles.

Waymo's safety framework is further supported by three basic types of system-level testing, which are computer simulation, closed-course tests, and on public roads. These three types of system level testing is supplemented by additional component and subsystem testing. Waymo also classifies three layers of the technology for its vehicles, including hardware, ADS behavior and vehicle operations. Here's a brief summary of each.

The Hardware Layer

The hardware layer includes the vehicle itself, beginning with Waymo purchasing fully certified vehicles from experienced vehicle manufacturers. These vehicles include the Chrysler Pacifica minivans and electric Jaguar I-PACE SUVs. 

After purchasing the vehicles, Waymo adds redundant braking and steering systems, which the company believes is a requirement for self-driving vehicles. The steering and braking components have robust technical requirements, including their performance in both normal and faulted conditions. 

Thes performance requirements are further backed by extensive testing and verification, including hardware-in-the-loop, closed course testing and evaluation in the field.

In addition to the base motion control systems that are added to Waymo's vehicles, the company installs a comprehensive suite of sensors that support autonomous driving, including  lidar, radar, cameras. All of the sensing systems are designed to meet rigorously defined performance and safety requirements, Waymo wrote in its report. 

The vehicle software that controls fully- automated driving also undergoes rigorous testing. Waymo said it developed a state-of-the-art computational platform that combines extreme performance with proven reliability and fault tolerance. The platform is designed with fault responses designed to address any point of failure. 

Finally to protect the vehicles from cyberattacks, Waymo also developed a process to identify, prioritize, and mitigate any threats from outside. Waymo joined the Automotive Information Sharing and Analysis Center's (Auto-ISAC), which is an industry-operated initiative created to enhance cybersecurity awareness and collaboration across the global auto industry.

The Behavioral Layer

The behavioral layer describes the software that safely operates the vehicles on public roads. Waymo uses three criteria for evaluating the performance of the behavioral layer, which are the avoidance of crashes, the successful completion of trips in autonomous mode and adhering to all traffic laws.

Waymo then performs scenario-based verification on its software to ensure that the ADS behavior meets all of the requirements of its "Waymo Driver." Next the ADS is tested using large scale simulated deployments where its further evaluated it based on the company's own rigorous performance metrics.

The Operations Layer

The operation layer for Waymo's vehicles includes the adherence to industry-leading safety practices for their safe operation. This includes a fatigue management program for Waymo's trained vehicle operators that sit behind the wheel to monitor the vehicles for testing, as well as details like the use of seat belts by all vehicle occupants.

The operations layer also includes plans for emergency responders, police and other law enforcement agencies on how to deal safely with driverless vehicles in case of an incident. Waymo also has its own fleet response team that can provide assistance to the ADS over a cellular network if needed, including taking control of a Waymo AV remotely.

Waymo's risk management team identifies, prioritizes, and works to resolve any safety issues before new or updated features or software are used on public roads. 

Waymo also has a dedicated safety program that identifies and addresses any potential safety issues based on information collected after any new software is deployed on public roads. 

The field safety process collects and helps resolve potential safety concerns from many other sources, including employees, riders, the public or Waymo suppliers.

"Waymo will reduce traffic injuries and fatalities by driving safely and responsibly, and will carefully manage risk as we scale our operations. This philosophy provides the foundation for all of our activities, is consistent with having safety at the center of Waymo's corporate culture, and sets the course for Waymo's future," the company wrote in its report.

Waymo based all of its comprehensive safety requirements with the goal of reducing traffic injuries and fatalities by driving safely and responsibly, and to manage risks as the company expands its operations. 

The company said it will continue to apply and adapt those methodologies to build an ever safer and more capable autonomous vehicle in the future. 

The publication of the two reports follows Waymo opening its fully autonomous Waymo One ride-hailing service to the general public in Phoenix.

FutureCar Staff