Note: Woven Planet became Woven by Toyota on April 1, 2023.
By Andrew Calkins, Senior Solution Engineer
Our goal at Woven Planet Holdings, Inc. (Woven Planet) is to benefit humanity by enabling the safe mobility of the world — “Mobility to Love, Safety to Live.” Within the Automated Mapping Platform (AMP) team at Woven Alpha, Inc. (Woven Alpha), we know that to make a big impact on a global scale, we need to work together with as many companies and solutions as possible. Working together we can achieve a greater good than is possible by going at it all alone. A key struggle to work with solutions from companies scattered throughout the globe, is that all those different technologies need to work together. The technology needs to speak the same language in order to achieve this cooperative advantage, and we can achieve that by working through standards.
In the high-definition (HD) Map industry, one such mapping standard leading the mapping world for the last 20 years is the Navigation Data Standard (NDS). NDS has been integrated with countless vendors and handled complex real world modeling scenarios across the globe. But as typical with standards that have been around for a long time, it is difficult to keep them updated. They can carry the baggage of what made sense in the technology world of the past but no longer makes sense today or in the near future. One such example is storing map attributes on a navigation road network instead of the detailed HD lane network. In the past, NDS.Classic attributes were sometimes forced to be stored on the less accurate navigation network. This required sending the map navigation network and the lane HD network both to handle certain cases. The additional challenge was that the Navigation network and the HD map network were fundamentally different in accuracy and even update frequency. The mapping table of these two fundamentally different maps was difficult to keep in sync. NDS.Live on the other hand has a flexible approach that allows assigning different road characteristics to the lane network. By assigning attributes to the lane network rather than the road network, this approach reduces the size of the overall payload (there is no need to send a road network). In addition, assigning attribution directly to the HD lane network significantly simplifies the attribution modeling. These simplications are exactly why the NDS consortium created the new standard called NDS.Live. With the expertise gained from NDS’ long market experience, the NDS.Live specification was created by bringing together the expertise from a wide range of mapping companies, system vendors and automotive manufacturers.
Here at Woven Alpha, creating an HD map product that can be used by many suppliers and providers is essential to our vision. At the same time, NDS.Live is not the only useful mapping standard out there. Some developers may want to work with other formats such as OpenDrive, or just consume data as-is in GeoJSON format. In the future, we may even develop our own format if developers find that existing formats don’t support their use case. Ultimately the developer community and our customers will influence what formats we will support to help meet their requirements.
Figure 1. NDS.Live sample Japan HD Map layers visualized through NDS MapViewer tool.
The advantages of NDS.Live Smart Layer and Registry
At Woven Alpha we recognize the value in the phrase “Build once, sell often.” The desire is to have one set of software that satisfies the needs of as many customers and markets as possible and this is particularly important in the mapping arena. With a wide variety of different customers and application use cases, it is difficult to find a common set of data that fits all customer needs. In addition, data transmitted to vehicles has to be optimized for size so as not to incur large communication costs. The NDS consortium set out to solve this problem by creating the NDS.Live Smart layers. The idea behind a smart layer is that it can be configured to deliver only the data layers that are needed by an individual customer’s use case, without delivering the entire data model. This allows for common software packages to be built to consume data layers targeted for specific use cases — For example, a curve warning application that needs lane layer, and Advanced Driver Assistance Systems (ADAS) curvature data, does not need to consume the same amount of map features as a Lane Keeping Assist application that uses localization data and detailed boundary/road surface marking data. By using Smart Layers and the registry, the different applications can minimize data transfer to only the data needed without having to create customized formats for each customer’s use case. In addition, the Smart Layers are self describing so that it can essentially tell the client application how to decode the received data layers. Coupled with the registry service, where customers can search for and request nodes that handle their particular use case, one set of software can now manage multiple client needs.
By embracing mapping standards like NDS.Live, AMP can efficiently and economically deliver our map data at scale. This in turn enables more customers to use our products. Further allowing us to help deliver even lower cost maps to more customers, and enabling developers to develop cool applications that will enable everyday safe automated mobility anywhere on the planet.
Join us! 👋
We are actively hiring exceptional talent for the AMP team. If you are interested in working with us in our brand-new office in Nihonbashi, Tokyo, please have a look at our current open positions!