GeoMate: AI-Powered Mapping for Smart Mobility
HD MAPS FOR AUTONOMOUS VEHICLES
Autonomous vehicles (AVs) promise to bring a wide range of benefits in the near future. These include everything from reduced accidents and traffic congestion to enhanced transportation accessibility. However, in order to deliver on these promises, self-driving cars need to be able to operate in complex environments more safely and effectively than human drivers. To do this, AV and Advanced Driver Assistance Systems (ADAS) developers use a variety of technologies. These innovations help vehicles understand and move in the world around them, including critical inputs like high definition (HD) maps.
WHY USE HD MAPS FOR AUTONOMOUS DRIVING?
HD maps are highly detailed, highly accurate maps that represent location data. Most vitally, they include features not normally found on traditional maps (i.e., lane markings, road edges, signage, etc.). HD maps help autonomous vehicles figure out what the world around them looks like. That ability supports a variety of key operational functions, including localization, perception, navigation and control.
While HD maps are important to autonomous vehicles, there are some technical challenges that persist in both building and maintaining them. HD maps are typically developed using a combination of different data sources. These inputs include Global Navigation Satellite Systems (GNSS), radar, IMUs, cameras, and LiDAR. However, each source has its own unique shortcomings. This can create errors that aggregate and ultimately impact the accuracy, cost, and time required to create HD maps. For example, LiDAR produces a highly detailed 3D construction of the environment. However, LiDAR is incredibly expensive and time-consuming. It also falls short because of limited operating altitudes and lack of reliability in poor weather conditions.
HD MAPPING: A COMPETITIVE EDGE
GeoMate’s HD maps have some important differences from other HD maps used today. Specifically, our maps contain unique datasets, they provide an incredibly favourable cost-to-accuracy ratio, and they are created through novel & innovative methods.
1. Cutting-edge mapping methods. GeoMate builds its maps by combining a variety of high-resolution geospatial data. We process them through machine vision and machine learning algorithms to produce “digital twins” of urban environments. As a result, our maps are faster to build, update, and scale. This is because they use data sources that can be analyzed quickly and do not require fieldwork to map and remap operational environments.
2. Comprehensive static & dynamic data. Our maps contain a wide range of datasets of both static and dynamic variables in the environment. These impact vehicle navigation and localization, including datasets on sidewalks, road edges, lane markings, etc., as well as dynamic variables, such as pedestrian, vehicle, and bicycle traffic flows, sidewalk closures, weather conditions, and predictive collision risk scores. These datasets are often missed in HD maps today. This means that users have an incomplete picture of the urban mobility environment.
3. High accuracy at low cost. GeoMate’s maps maintain centimetre-level accuracy without the high costs of traditional HD mapping methods. For linear features (sidewalks, road edges, lane markings, etc.) our maps produce a positioning accuracy of ~3cm. For vertical features/urban assets (signage, lighting, fire hydrants, traffic lights, etc.), we provide positioning accuracy between 10-20cm. Because we rely solely on geospatial imaging rather than other inputs like LiDAR, we can keep the costs of our maps low while still maintaining the high levels of accuracy required for HD maps in AV/ADAS.
MAXIMIZE YOUR IMPACT WITH HD MAPS
Urban areas are becoming increasingly complex and dynamic spaces. As such, mapping these environments accurately is more important than ever for self-driving technology. At GeoMate, we are creating maps that keep pace with these changes.
We’re driving the future of HD mapping by re-thinking how we collect, analyze, and use geospatial data. In doing so, our service makes it easier, faster, and cheaper for our clients to map and scale their technology and create safer roads for everyone.
