The specific section of Montreal featured in our digital twin represents a variety of commonly occurring North American grid road structures and encompasses older, downtown areas. The many features associated with an urban environment can be found in abundance right here within this one model.
There’s a plethora of variety and transition within this digital twin. Most of the drive-able sections have a classic structure: building, sidewalk, park zone, road and boulevard. The model also includes four, three and two lane roads, both dual and single way, as well as elevation changes and a tunnel.
AV systems can be challenged by the presence of other road users such as cyclists travelling on the dual bicycle routes which insect the roads and pedestrians moving through the busy streets lined with shops and homes. Given the blend of large buildings, road side objects and tree lined streets there are plenty of opportunities for casting shadows when testing most perception systems. Since the time of day and weather effects can so easily be changed, this digital twin offers countless opportunities to find unique experiments and edge cases.
There is a unique slotted overhead structure in the city which can create a strobe effect across Montreal’s streets and roads. When the environment is set to sunshine, the shadows from this structure pose a real challenge for camera sensors. Furthermore, features as common as bicycle routes which feature heavily in this model, can cause confusion to perception algorithms, so they can be tested in abundance within Montreal.
This model provides a typical North American, downtown, urban setting, providing an extremely accurate simulator for drivers, cameras, sensors and algorithms to be safely tested in prior to encountering such conditions in the real world.