Transurban’s Victorian CAV trials: Phase 1 report

Over a period of four months Transurban, in association with the Victorian Government, VicRoads, and the RACV, have been conducting a trial of Connected and Automated Vehicles (CAVs). Phase one was a study of partially automated vehicles, and its findings have just been released.

Why is Transurban running these trials?

We own and operate roads in both Australia and the United States and it’s up to us to make sure our roads are ready for whatever changes we can see coming, whether it’s more people using our roads, changing freight patterns or the arrival of new transport technology.

To prepare for the expected influx of connected and automated vehicles on our roads, we’ve been running trials to investigate how they respond to our road infrastructure (such as signs, lines, signals and ramps).

The cars and technologies trialled

The testing took place on Melbourne’s Monash, CityLink, and Tullamarine motorways, between August and November 2017. Twelve cars from six manufacturers were used, namely:

• Audi
• BMW
• Mazda
• Mercedes-Benz
• Tesla
• Volvo

Each was a current model vehicle, with partial automation features — including Lane Keep Assist, Adaptive Cruise Control, Traffic Sign Recognition, and Minimal Risk Condition (what a vehicle does when a driver does not react/conform to multiple warnings to take back control of the vehicle from automated mode).

All up there were 4 sets of trials, 46 trial sessions, 118 on-road hours, and 4,900 kilometres travelled. A professional driver was behind the wheel, along with an observer, and four cameras provided video footage from each vehicle’s front, rear, and interior.

RELATED ARTICLE  How ready is Australia for autonomous vehicles?

Main findings

• the vehicles did not experience as much difficulty as expected in cases of temporary yellow lane marking used in tandem with the regular white lane marking
• gaps in lane marking underneath the electronic toll gantries caused the lane keeping systems of some vehicles to disengage
• changes in line marking (changes in line type from solid to dotted, added lane for merging from on-ramp, gaps caused by road expansion joints, drains, etc.) also caused lane keeping automation systems to disengage
• vehicles sometimes read and reacted to static road signs on exit ramps (when not actually on the exit ramp)
• electronic road signs were more challenging than static signs for some of the test vehicles, particularly in tunnels
• The CityLink’s Sound Tube art installation (pictured above) caused a few issues: disengagement of lane keeping for some vehicles; in one instance a car’s system identified a ‘ghost vehicle’ where there was no lane; and it caused some vehicles to detect a 110 km/h or ‘de-restricted’ speed limit
• Objects on the road, such as stopped vehicles, people exiting stopped vehicles, traffic cones, and temporary road signs, were not detected

In addition to publishing the findings from this trial, Transurban will from April 2018 begin to act on recommendations outlined in the report. This includes changes to the roads, and discussions with vehicle manufacturers.

Download the report

Click the button below for a full copy of Transurban’s ‘Victorian connected and automated vehicle trial, Phase One — Partially automated vehicles’.

View footage from the trials on Transurban’s Partially automated vehicle trials page.