BMW and Mobileye to develop crowd-sourced sensor data for automated driving
BMW Group and Mobileye have agreed to bring Mobileye's Road Experience Management (REM) data-generation technology to new group vehicles entering the market next year. The resulting crowd-sourced real-time data, using vehicles with camera-based Advanced Driver Assist System technology, along with next-generation high-definition (HD) maps, is a critical enabler for autonomous driving.
Inclusiveness and industry collaboration is a goal of the program to further promote automated driving. BMW Group sensor data would be merged with data from other automakers to strengthen Mobileye's Global RoadBook (GLRB) offering, which supports and rapidly updates HD maps with highly accurate localization capabilities. Autonomous vehicles will require HD maps that can identify and update changes in the environment in near real-time to enable short "time to reflect reality."
The cameras collect anonymized fleet-wide data, and Mobileye’s EyeQ processors and software identifies valuable information that is sent to the cloud in a highly compressed 10-kB/km form. The data can be used to add a dynamic layer to navigation maps, enabling BMW Group customers to access true real-time information on traffic density, potential road hazards, weather conditions, on-street parking, and other information.
The BMW Group and Mobileye will transfer anonymized data to Here, the mapping and location service company, which will use the data and information to conduct real-time updates of its HD Live Map, the real-time cloud service for partially, highly, and fully automated vehicles, and enhance its Open Location Platform—all ensuring an accurate depiction of the driving environment in real-time. Mobileye and Here earlier announced their intent to integrate data gathered through REM technology as a layer in HD Live Map.
The BMW Group stresses that it is open to collaboration with additional partners, including OEMs or other third parties. This Mobileye announcement underlines its inclusive approach to automated driving cooperation. The approach includes its development with Intel and Mobileye on bringing highly automated driving to market by 2021 with iNEXT. The BMW Group believes it brings to any collaboration its automotive competence in safety and software standards, motion control, and end-to-end system integration.
"This announcement demonstrates that our partnership is rapidly bringing innovation to market to allow customers to benefit from the latest technology," said Klaus Fröhlich, Member of the Board of Management of BMW AG for Development. "At a strategic level, this announcement makes it clear how our cooperation with Mobileye leverages our investment stake in Here. The data of future BMW vehicles will enrich the quality of maps and services for everyone. Furthermore, this represents a significant step toward introducing the BMW iNEXT, with its features of highly automated driving in 2021, as well as creating a leading ecosystem around Here's Open Location Platform leveraged by swarm data from millions of vehicles across the world."
"We welcome the opportunity to take this next step in our relationship with BMW Group," stated Professor Amnon Shashua, Chairman and Chief Technology Officer of Mobileye. "Camera-based ADAS systems are already making the roads safer. Global RoadBook is an initiative to utilize data from these cameras to create the high-definition maps required to make the next generation of autonomous driving a reality, in an inclusive way which will create an industry standard."
THE POWER OF THE CROWD
The need for HD maps to enable fully autonomous driving would satisfy functional safety standards that require back-up sensors, or redundancy, for all elements of the chain—from sensing to actuation. This applies to all four elements requiring sensing: free space, driving paths, moving objects, and scene semantics.
While other sensors such as radar and LiDAR may provide redundancy for object detection, the camera is the only real-time sensor for driving path geometry and other static scene semantics (such as traffic signs, on-road markings, etc.), according to Mobileye engineers. Therefore, for path sensing and foresight purposes, only a highly accurate map can serve as the source of redundancy. For a map to be a reliable source of redundancy, it must be updated with an ultra-high refresh rate to secure its low time-to-reflect reality (TTRR) qualities.
Mobileye is addressing this challenge by harnessing the power of the crowd: exploiting the proliferation of camera-based ADAS systems to build and maintain in near-real-time an accurate map of the environment. Its REM end-to-end mapping and localization engine for full autonomy is comprised of three layers: harvesting agents (any camera-equipped vehicle), map aggregating server (cloud), and map-consuming agents (autonomous vehicle).
Harvesting agents collect and transmit data about the driving path’s geometry and stationary landmarks around it. Mobileye’s real-time geometrical and semantic analysis, implemented in the harvesting agent, allows it to compress the map-relevant information into a very small communication bandwidth of less than 10 KB/km on average. The relevant data is packed into small capsules called Road Segment Data (RSD) and sent to the cloud. The cloud server aggregates and reconciles the continuous stream of RSDs, a process resulting in a highly accurate and low TTRR map called a Roadbook.
The last link in the mapping chain is localization. For any map to be useful, an autonomous vehicle must be able to localize itself within it. Mobileye software automatically localizes the vehicle within the Roadbook by real-time detection of all landmarks stored in it.
REM provides the technical and commercial conduit for cross-industry information sharing. It is designed to allow different OEMs to take part in the construction of the Roadbook automated-driving-critical asset while receiving adequate and proportionate compensation for their RSD contributions.