The latest software releases have rolled out to the driverless fleet, containing improvements in these areas:
Shipped TSEL v20.4, increasing politeness and caution while yielding to pedestrians in crosswalks, and while interacting with other vehicles driving aggressively or recklessly
Launched TSEL v19 which improved the AV's response to other drivers cutting-in front, continuing navigation while keeping a safe distance from such vehicles; improved AV's behavior around birds and debris in the road; and improved passenger pickup and dropoff by navigating closer to curb for out of lane pullovers by 38%.
Improved maneuverability through narrower gaps, tight turns, and around debris, resulting in smoother driving.
Launched LPC v11, increasing ride comfort by improving how the AV maneuvers through narrow gaps, while enabling the AV to dynamically adjust driving behavior based on road grade.
Enabled the AV to dynamically adjust its on-road response to wet roads when there is no active rainfall, resulting in smoother driving in such weather conditions.
Launched Oseg v5 and Semantic Grid v1, improving the AV's autonomous detection of closed roads and response to objects such as caution tape and fire hoses.
Improved AV's response to road closures and EMVs by rerouting for blockages and keeping the intersection clear at road closures.
Improved the AV’s navigation near construction sites during Remote Assistance.
Shipped VRU MTL v3.1 that improved the AV’s ability to detect gestures from humans directing traffic by another 5%.
Launched Cyclops v1 that improves perception of nearfield objects by 24% and far away high-speed vehicles by 20-30%.
Shipped DLA v7 and KSE-Vehicles v18 that improve tracking of long range and high speed vehicles by up to 12%.
Further improved perception of overhanging tree branches and other inanimate objects above the AV, allowing for a smoother ride with less temporary trip interruptions.
Shipped X-MTL v2 along with motion control improvements for smoother driving around small debris and birds.
Improved passenger pickup and dropoff by minimizing delays when passengers enter and exit the AV.
Improved Pick-up and Drop-off experience by safely crossing bike lanes to stop in parking spots
Increased reliability of the AV stopping at the intended destination during pickup and drop-off on the first try.
Expanded the AV's ability to resume autonomously after a trip interruption.
Expanded the AV's ability to remotely recover when facing multiple, simultaneous recovery conditions.
Expanded the ability for our AVs to remotely recover from tampering events.
Reduced trip interruptions through improved localization reliability while stationary.
Continued expansion of AV ability to maneuver and park in our charging facilities, keeping our AVs charged up and on the road.
Statistics measuring the percentage of improvement are based on internal benchmarks and metrics.