The level of risk minimization offered by automated driving relies on data collected by HD maps. We outline how companies and industries collaborate to drive the best results.
Human error accounts for nearly 98% of traffic accidents in Germany.1 As interest in automated driving technology has expanded, driven in part by the initiative to lower vehicle-related injuries and fatalities, demand for safer automated vehicles grows ever more relevant. Certainly, automated driving technology offers potential for far safer transport in the future.
However, the full benefits of automated driving depend heavily on the reliability of data collected by its onboard and off-board sensors, and onboard sensor capability alone is inadequate to meet the safety requirements of automated driving. As automated technology continues to progress, intelligent HD map perception has never been more important in guaranteeing the safety of automated driving operations. We, the NDS Association members, believe that maps are a necessary complementary reference to enable the processing of important information to detect road features inaccessible to on-board sensors.
Intelligent HD maps are utilized in circumstances where on-board sensors cannot detect certain road features, such as road boundaries or traffic guidance. In-vehicle maps also provide location-based determination, environment modeling in adverse conditions and precise semantic understandings in complex driving situations. As such, a priority for developers, manufacturers and operators alike is to guarantee intelligent HD maps, alongside onboard sensors, facilitate a collision free and lawful driving plan for automated vehicles. Maps also enable an electronic horizon far farther than on-board sensors can, which allow a high level of comfort and feeling of safety that is required for consumer adoption of automated driving.
This year, leading authors and experts from 11 industry leaders – Aptiv, Audi, Baidu, BMW, Continental, Daimler, FCA US LLC, HERE, Infineon, Intel and Volkswagen – published Safety First for Automated Driving (SaFAD), the largest report to date on development, testing and validation of safe automated passenger vehicles. The SaFAD whitepaper provides automated vehicle developers and operators a system outlining how automated vehicles can prove to be safer than human drivers by streamlining and utilizing technology.
The SaFAD initiative and NDS Association encourage collaboration across automakers, suppliers, and technology industries. For example, announced at this year’s Open Auto Drive Forum event, fellow NDS Association members Elektrobit (EB) and Baidu, Inc. (NASDAQ:BIDU) signed an agreement whereby EB will provide automotive infrastructure software for Baidu’s Apollo Computing Unit (ACU). The strategic partnership, outlined by a whitepaper on reliable and safe maps for automated driving, ensures HD maps have the best dependability representation of reality required for safe navigation.
Apollo is Baidu’s in-vehicle computing platform for automated driving. Apollo’s localization system is a comprehensive positioning solution with centimeter level accuracy based on GPS, IMU, HD map, and a variety of sensor inputs capable of generating large scale HD mapping data. EB’s software, EB tresos, enables the development of AUTOSAR standard-compliant software for automotive electronic control units (ECUs) with the highest levels of functional safety.
Now paired with EB tresos, Apollo Computing Unit extends the application of deep learning and artificial intelligence technology to map creation, but ensures automated vehicles are equipped with intelligent HD maps for safe transport. In a world of ever increasing interconnectivity, automated vehicles require stream lined systems with common standards to provide the highest level of efficiency, reliability and security for their future passengers.
1. German traffic accident statistics, published by the Federal Statistical Office of Germany (Destatis, 2018: Fachserie 8, Reihe 7, Verkehr, Verkehrsunfälle, 2017, Germany) ↩Back to news →