GPR’s Ground Positioning Radar™, like traditional sensors, uses landmarks to position autonomous vehicles relative to road features that matter. The key difference is that GPR landmarks are deep beneath the driving surface. This character-rich data, which is extremely unique even in seemingly homogeneous environments, is the key to reliable and accurate localization across terrains.
GPR maps extract subterranean landmarks to provide exceptionally robust vehicle positioning. Fundamentally, GPR operates like any other Landmark Localization system, but with the distinct advantage of a consistently detectable, stable data layer. Since GPR data is protected by the ground, it is always available, rarely changes, and is highly unique.
The Missing Modality
Independently, or in combination with other sensors, GPR provides mission-critical localization for everything that moves autonomously, down to the centimeter.
Seeing What Other Sensors Can't
Even the most advanced sensors fail to provide a reliable position in many scenarios. Current localization technologies, including lidar and camera-based systems, rely on an uninterrupted line of sight to detect the relevant landmarks.
Line of sight can be easily obstructed by surrounding objects like other vehicles or harsh conditions like snow, heavy rain, and fog. Mud, snow or ice can cover traditional sensors causing partial or full blockage events. In addition, the surface landmarks traditional localization systems rely on can be homogeneous, scarce, or frequently changing making them unreliable and poorly suited for localization.
The Surface Changes. GPR Maps Don’t.
By comparison, the subsurface is exceptionally stable, character-rich, and accessible.
Mapping is easily achieved via crowdsourcing or mapping fleets. Any vehicle equipped with GPR for localization can expand, update, or validate the GPR map layer.