In recent years, the landscape of automotive mapping and navigation has undergone a paradigm shift, driven by significant technological advancements and an increasing demand for smarter, more efficient transportation solutions. These developments pave the way for enhanced driver assistance, intelligent e-mobility solutions, and the advent of autonomous driving. As a leading provider in software engineering for the mobility industry, Intellias is significantly contributing to these innovations.
Despite improvements in obtaining mapping and navigation data through tools like dashcams, UAVs, and satellites, data collection remains labor-intensive. Even with extensive mapping coverage, maintaining data accuracy and timeliness stands as a persistent challenge due to the ever-evolving nature of the physical world. To address these demands, the evolution in navigation and digital mapping technologies is accelerating, driven by six major trends.
AI Integration in Mapping Data
Enhancing Speed and Accuracy with AI
Satellite imagery revolutionized map creation; however, traditional mapping software cannot directly utilize satellite photos. The visual data must be coded into usable formats such as the Navigation Data Standard (NDS), making the map creation process both costly and labor-intensive. Here, artificial intelligence (AI) plays a transformative role by enhancing the speed and accuracy of digital mapping. AI algorithms enable more frequent updates of maps and faster mapping of new areas. For instance, AI can classify objects in satellite images to create enriched two-dimensional digital maps and multilayer three-dimensional models. Accurate maps result in improved estimated times of arrival (ETAs), better fuel or energy usage estimates, and more detailed points of interest.
In addition to aiding data collection, AI can generate mapping data. MIT and Qatar Computing Research Institute developed RoadTagger, a neural network that predicts road types and lane numbers even when visual obstructions such as trees or buildings are present. This enhances digital map accuracy. Also, with the rise of connected vehicles, Original Equipment Manufacturers (OEMs) increasingly rely on fleet data for map creation. For example, HERE Technologies introduced UniMap, an AI-driven solution for rapid sensor data processing and map creation, enabling new digital maps within 24 hours.
NDS.Live: Transitioning from Offline Databases to Distributed Map Data Systems
The Shift to Hybrid and Online Navigation
Traditional onboard navigation systems rely on proprietary databases that quickly become outdated. NDS.Live is the new global standard for map data, promoting the shift from offline to hybrid and online navigation. By minimizing the complexity of supporting various data models, storage formats, and protocols, NDS.Live offers a flexible specification. It is not merely a database but a distributed system of map data. It was co-developed by prominent global OEMs and tech firms including Intellias, Daimler, HERE, Denso, Renault, and TomTom.
Real-Time Driver Assistance with NDS.Live
For instance, the second-generation Mercedes-Benz User Experience (MBUX) system uses NDS.Live to provide real-time driver assistance visualized as augmented reality (AR) on the head-up display (HUD). This system significantly enhances navigation for electric vehicles (EVs) and regular connected cars, supporting OEMs in deploying value-added subscriptions for assisted driving and navigation.
3D and HD Map Generation
Importance of 3D and HD Maps
Three-dimensional (3D) maps are pivotal for accurately rendering physical objects, while high-definition (HD) maps include detailed information on road features and terrain types. Both map types are vital for advanced driver assistance systems (ADAS) and autonomous driving. 3D maps facilitate vehicle movement interpretation, assisting navigation systems with additional road and terrain information, which is especially crucial for the limited range of onboard sensors.
Industry Collaboration for HD and 3D Maps
However, collecting and rendering data for HD and 3D maps is challenging, requiring vast resources and collaboration across the industry. For instance, HERE and Mobileye, with the later inclusion of Volkswagen (VW), have partnered to crowdsource HD mapping data. Mobileye provides a compact vision system, EyeQ, which feeds data into HD maps rendered through partnerships. Similarly, TomTom collaborates with Qualcomm Technologies to gather HD map insights from its users via a cloud-based platform, further enhancing the navigation data.
Autonomous Driving Simulations
Virtual Testbeds for AV Development
Extensive road and track tests, including the virtual simulation of near-crash events, are essential for autonomous vehicle (AV) development. Hyper-realistic virtual worlds now serve as safer, controlled testbeds for AVs. Researchers have developed an open-source simulation engine to create photorealistic training environments for AVs, capable of simulating complex sensor data and driving scenarios.
Real-World Data for Simulation
Companies like Waymo use real-world data to build detailed virtual replicas of driving conditions, allowing their algorithms to master challenging driving interactions through repeated simulations in various conditions. The high-fidelity 3D environments effectively convey real-world details, enhancing ADAS/AD scenario simulations close to actual conditions.
Digital Twins of Road Infrastructure
Benefits of Digital Twins
Dashcam data collection now also benefits transportation management for digitizing road infrastructure. Digital twins are virtual representations of physical systems like smart traffic networks or parking facilities, powered by real-time data. They offer advanced urban planning capabilities, optimizing traffic lights, transport management, and accurate traffic predictions. Low latency in 3D map generation at the edge is crucial, warranting substantial computing power. Researchers propose compact roadside distribution devices for real-time point cloud data map delivery, enhancing autonomous vehicle localization and navigation efficiency without storing extensive data onboard.
AR in HUD Navigation Products
Benefits and Future Developments of AR in HUDs
HUDs in modern vehicles have advanced to incorporate AR, displaying dynamic driving instructions and real-time route details directly in the driver’s field of view. AR navigation systems ensure drivers make better decisions, reduce errors, and can improve driving performance and comprehension. Future developments include holographic displays with 3D AR instructions, using advanced lidar technologies to project ultra-HD road object representations in real-time. This innovation could shorten obstacle visualization times and alleviate driving-related stress.
Enhancing Driver Experience with AR
AI isn’t just revolutionizing data collection; it’s also transforming map generation. MIT and Qatar Computing Research Institute developed RoadTagger, which is a neural network designed to predict road types and lane numbers even when there are visual obstructions like trees or buildings. This innovation significantly boosts the accuracy of digital maps. Additionally, with the surge in connected vehicles, Original Equipment Manufacturers (OEMs) are increasingly relying on fleet data to create maps.
A great example is HERE Technologies, which launched UniMap, an AI-driven solution designed to speed up sensor data processing and map creation. This tool can produce new digital maps within just 24 hours, showcasing the incredible advancements in this field. AI’s ability to enhance mapping goes beyond mere road types and lane numbers. It integrates data from various sources to offer comprehensive, real-time maps. This technology can adapt to changes swiftly, ensuring that maps are always up-to-date.
Moreover, the collaboration between AI and connected vehicles is paving the way for even more precise and dynamic maps. OEMs can utilize the constant stream of data from these vehicles to update maps continuously, improving navigation systems, and enhancing driver safety. In short, AI is playing a pivotal role in advancing mapping technologies, making them more accurate, efficient, and reliable than ever before. With ongoing research and development, the future of digital mapping looks incredibly promising.
