The Fuse

This Week in AVs: GM, Ford & Toyota Advance Self-Driving Testing, Standards; Daimler Trucks Takes Control of U.S. AV Tech Group; and More

by Kristen Hernandez | April 08, 2019

GM, Ford and Toyota join to advance self-driving testing, standards
Three major automakers are teaming up with SAE International to develop autonomous vehicle safety standards that could influence future regulations. The Automated Vehicle Safety Consortium, which includes General Motors, Ford, and Toyota, intends to work with both industry and government to pursue this objective. They will initially focus on prioritization, currently emphasizing data sharing, interactions between autonomous vehicles and other road users, and guidelines for safe AV testing. While the current patchwork of regulations has led to varying autonomous vehicle testing requirements and the public remains uncomfortable with the technology, developing standards could facilitate development of federal regulations and assure the public of AV safety.

Aptiv’s data set gives users access to 1.4 million images and 390,000 lidar sweeps, as well as radar, accelerometer, and GPS data.

Aptiv Releases Comprehensive Open-Source Automated Vehicle Data Set
Aptiv, a leading automotive supplier, announced the release of nuScenes by Aptiv, an open-source data set collected from autonomous vehicles in Singapore and Boston. Aptiv’s data set gives users access to 1.4 million images and 390,000 lidar sweeps, as well as radar, accelerometer, and GPS data. The company claims their data set includes a variety of scenarios that can help users train their systems on the challenges of both right- and left-hand urban driving environments. Given the time-consuming processes of data collection and labelling, data sets such as this could significantly cut down on the time it takes autonomous vehicles to learn new skills and a new environment. This could lead to accelerated deployment across a variety of settings. 

Daimler Trucks takes control of leading US self-driving tech group
Late last week, Daimler Trucks, the world’s largest producer of commercial vehicles, bought a majority stake in self-driving tech company Torc Robotics. This purchase follows the company’s promise in January of spending €500 million on accelerating autonomous vehicle technology and marks one of the first times a European company has taken over a US autonomous vehicle company. Torc’s unique achievement of having been profitable every year since its founding may be part of what drew Daimler to the company. Historically, most of Torc’s revenue has come from defense work.

Search giant Baidu has driven the most autonomous miles in Beijing
Beijing’s first report on the city’s licensed self-driving cars revealed that Baidu has a significant lead in distance driven. Out of the roughly 150,000 kilometers driven by eight firms in the city, Baidu accumulated almost 140,000, about 91 percent. Pony.ai, the runner-up, drove only 10,000 kilometers. While this is the first glimpse the public has been given of data on self-driving testing in Beijing, it is relatively uninformative. Beijing only reported kilometers driven, forgoing even the (flawed) disengagement rates reported by California (Baidu and Pony.ai also test their vehicles in the Golden State). These metrics lack any insight into the complexity of conditions under which the vehicles were operating. Until both Beijing and California can offer more insight on the conditions behind this data, the data provided offers little information on the progress being made by autonomous vehicle developers.