Optimising HD Map Creation Within Tunnels

Project objective

Test HD map creation in indoor scenarios

Location

Nanjing, China

Operation site

Tunnel

Vehicle

Autowise V12 and Autowise V16

Autonomous service providers have often found that data can be collected with relevant ease in an open area. However, in an underground or enclosed area with weak network signals, there is a time delay that affects the transfer of data from lidars to the computing system. Additionally, in an enclosed area, GNSS cannot function. As such, engineers have worked tirelessly to develop solutions that allow an autonomous vehicle to create HD maps and identify its location as it transitions from an outdoor area to an enclosed indoor area.

On March 28th, 2021, Autowise.ai successfully deployed two autonomous sweeper trucks in the Nanjing Yangtze River Crossing Tunnel. The tunnel is a 5853m long twin-tube that crosses the Yangtze River in Nanjing, China. On a regular day, the tunnel has approximately 80,000 vehicles and 190 trains driving through it at up to 80 km/h.

The new algorithms allowed the LIDAR to emit lasers with frequencies compatible with various network signals. The data substituted the missing GNSS positioning data by adding pinpoints in the HD map for vehicle identification.

The following user case is based on actual project results. For more information, do not hesistate to contact us at

[email protected]