Over the past year, we have seen an increase in the use of robot taxis and autonomous vehicles. Companies such as Waymo, Cruise, and Baidu have all made solid strides as industry pioneers. In China in particular, the headlines of 2020 regularly featured big announcements of autonomous vehicles, such as the public launch of the Baidu Apollo robotaxi services in the cities of Beijing, Changsha and Cangzhou.
But despite the increasing visibility of robotaxis and the wider public exposure to autonomous vehicle technology, many people remain hesitant about the safety of self-driving cars. Nearly three in four Americans say autonomous vehicle technology “is not ready for prime time,” according to a survey by Partners for Automated Vehicle Education (PAVE). Almost half (48%) say they never get in a taxi or autonomous carpooling vehicle.
Building trust between human passengers and self-driving cars is fundamental to the public’s use of self-driving vehicles. The current lack of confidence may stem from controversies and high-profile accidents involving passengers and autonomous vehicles. Tesla’s Full Self-Driving (FSD) vehicles have raised concerns as its technology was launched without proper safety-focused communication, while requiring drivers to constantly monitor the road. Uber’s autonomous vehicles sparked controversy after a test drive ended in fatality when the vehicle failed to recognize a pedestrian. The risks and mistrust of autonomous vehicles make safety the top priority for companies and their passengers.
As a leading player in autonomous driving in China, Baidu has made significant strides in building public confidence in its autonomous vehicle technology. Part of its success stems from the continued development of cutting-edge security features, such as a state-of-the-art artificial intelligence (AI) system and 5G compatible teleoperation. But equally important has been its deliberate focus on developing these cutting-edge security features and communicating how these technologies and best practices – such as information security and security assessments – ensure the travel safety by minimizing risks and accidents.
So what role do these technologies play in improving the safety of self-driving cars? How did Baidu communicate around these features to allay consumer concerns? And what can other companies in Baidu’s journey learn to better develop public confidence in autonomous vehicles?
A reliable AI system
Baidu Apollo’s fleet of nearly 500 self-driving vehicles is known for its reliability and performance. Self-driving cars have traveled more than 7 million kilometers (4.35 million miles) without an accident and have safely transported more than 210,000 passengers. In 2019, Baidu’s autonomous vehicles traveled 108,300 miles in California, with just six walkouts and no accidents or injuries. Likewise, Baidu’s 52 autonomous vehicles traveled nearly 468,513 miles in Beijing in 2019 without incident.
Mock tests and a well-designed verification mechanism are the key to building a safe autonomous vehicle. A 2018 report who demonstrates the reliability of autonomous vehicles calculates that it would take 8.8 billion miles – with a fleet of 100 vehicles tested 24 hours a day, 365 days a year, at an average speed of 25 miles per hour.
Simulation makes large-scale testing of autonomous vehicles virtually possible. Each iteration of the Baidu Apollo system is tested over millions of kilometers every day in a virtual environment that now accommodates more than 10 million simulation scenarios. Baidu researchers also presented a method to augment real world images with simulated traffic flow to create photorealistic simulation scenarios.
Before being tested on public roads, the system undergoes a series of hierarchy checks, from a fully virtual environment to mixed reality through closed areas. Every self-driving car module, including models, software, sensors, and vehicles, will be fully reviewed.
Baidu Apollo Integrated AI System enables its vehicles to drive autonomously without a safety driver inside the vehicle. To ensure public safety in extreme road conditions, Baidu has integrated 5G teleoperation into its vehicles.
The 5G remote driving service, powered by 5G networks, intelligent transportation systems and vehicle technologies for everything, provides immediate assistance to remote human operators during emergency situations. Operators aim to ensure the safety of passengers and pedestrians when its non-autonomous driving mode is used.
All remote human operators have completed more than 1000 hours of cloud driving training without any accidents and can ensure the safety of passengers and pedestrians when non-autonomous driving mode is activated. Since autonomous vehicle drivers can handle most road conditions, human intervention is a rare opportunity. Currently, Apollo’s autonomous cars in Changsha and Beijing have 5G compatible teleoperation that allows them to test on public roads without a safety driver.
No matter how effective the built-in AI systems are, systems like Baidu’s 5G remote driving service ensure that a person is always available to intervene.
To advance safety in mobility as a long-term vision towards developing an autonomous driving ecosystem, Baidu launched an automotive cybersecurity lab, becoming the only Chinese company to do so.
The lab studies automotive cybersecurity technologies, trends and solutions for in-vehicle systems, car-to-car communications, controller network, and sensors. It explores security best practices in data protection, in-car infotainment, reference software and hardware design, as well as countermeasures against false signals that deceive autonomous driving systems.
Knowing about vulnerabilities and potential cyber attacks are industry-wide concerns. In 2019, Baidu partnered with automotive industry leaders and released a white paper which described how to build, test and operate a safe automated vehicle. As part of its CTF Challenge competition, Baidu encouraged developers and white hat hackers to design protection mechanisms for autonomous driving systems.
The laboratory, broad information sharing and strong collaboration with industry are key elements in maintaining public confidence in autonomous vehicles. No system is completely immune to cyber attacks, as has been proven when researchers have demonstrated their ability to take remote control of a Jeep Cherokee (and technically most modern vehicles), controlling things like the steering, brakes and wipers.
Baidu’s focus on industry-wide security research and collaboration not only improves the security of these stand-alone systems, but also demonstrates that public safety is its top priority when vulnerabilities inevitably arise. exhibited.
Operations safety assessments
The final way Baidu continually assures the public that its autonomous vehicle systems are working effectively is through regular, in-depth safety assessments. Baidu Apollo has a full-time “safety assurance team” that aims to prevent robotaxi accidents. The team is responsible for performing various aspects of the safety assessments of the operating area, management of the operating platform, training of safety drivers and safety assurance mechanisms. . Team members assess the external environment of autonomous vehicles, including how to configure the road network and how to configure and expand the site.
Like any equipment with inherent safety risks, failures will occur and safety systems can always be improved. By constantly reviewing and revising the safety measures it implements in its autonomous vehicle deployments, Baidu can spot problems before they lead to an accident, while constantly reassuring the public that it is taking action. safety seriously.
Building confidence to move forward
Building trust between the public and autonomous vehicles is fundamental for their mass adoption. As an industry pioneer, Baidu has focused much of its innovation on developing and demonstrating technologies and best practices that can minimize risk. Through these initiatives, Baidu has developed a solid foundation of public trust in China that other companies can build on to move forward.
The key is that technological innovation is only half the battle. Reliable security features such as AI systems, teleoperation options, and enhanced information security are essential. But perhaps even more important is to communicate how these innovations not only improve safety, but compare to the everyday risks we accept with manual driving. It is the model for changing public sentiment and confidence in autonomous vehicles.
This content was produced by Baidu. It was not written by the editorial staff of MIT Technology Review.