JEREMY NASSAU, HEAD OF FUTURE TRANSPORT AND PARTNERSHIPS
TRANSURBAN
Results from Australia’s first trial involving a highly automated truck operating in live-traffic conditions are in. The trial, run by toll-road-operator Transurban, found that infrastructure to road (I2V) connectivity has potential in supporting the future transition to connected and automated vehicle (CAV) operations on Australian roads.
Autonomous trucks are attracting interest as one of the more feasible applications of automated vehicle technology – particularly as a solution for delivering safe and efficient freight movements (on suitable routes and under appropriate conditions). To this end, on-road automated heavy-vehicle pilots have been conducted in the US and Europe – and now Australia.
Many Transurban roads form a critical part of major urban freight corridors, so our automated truck trial helped us understand automated trucks’ potential in improving freight efficiency in Australian cities. Given road freight is projected to grow steadily in the coming decades,[1] identifying potential freight solutions, and mitigating barriers to their realisation, is something we need to be exploring now.
Further, as CAV development has progressed, the challenges of safely introducing these vehicles onto roads have become more evident across the industry. But while CAVs’ widespread adoption appears to be a way off, our trial findings suggest that introducing these vehicles in a controlled environment, delivered via smart infrastructure, is a potential path forward.
We ran our automated truck trial working closely with the Victorian Government’s Department of Transport and Planning (DTP) and with the Industrial technology Research Institute (ITRI), the truck’s Taiwan-based automated driving system (ADS) developer. Trials were run over four weeks in late 2022, along Transurban-operated CityLink and the DTP-operated Monash Freeway in Melbourne. Trials were conducted at night when traffic numbers are lower.
[1] Freight Australia, Navigating Australia's Freight Future, Navigating Australia's freight future | National Freight Data Hub (freightaustralia.gov.au), accessed 21 June 2023.
TRIAL TECHNOLOGY AND PREPARATIONS
AUTOMATED TRUCK SPECIFICATIONS
The survey provides important insights into the usage and desirability of driver assist functions.
The truck used in the trial was a left-hand-drive modified Hino 700 SH 4x2 tractor with built-in heavy-duty capability for towing a trailer with a shipping container. The full ADS kit (pictured), retrofitted to the vehicle by ITRI, included:
- two industrial computers and ADS monitoring equipment (installed in cab).
- a sensor suite comprising: mounted lidars; inertial measurement unit (IMU); cameras; and radar.
ON-ROAD ENVIROMENT
Transurban roads, including CityLink, are among the most high-tech roads in the world, with embedded technology such as CCTV cameras, automatic incident detection systems and smart sensors that monitor for debris, stopped vehicles and other traffic events.
This trial leveraged CityLink and the Monash Freeway’s technology and managed motorway environment and combined testing the highly automated truck’s operations alongside innovative digital solutions developed for the trial, including:
- integrating road-infrastructure data into the ADS, extending the automated truck’s awareness of road and traffic conditions beyond its sensor range – including helping it see around corners
- implementing a dynamic dedicated lane around the truck as it operated.
DYNAMIC DEDICATED LANE
To simplify the truck’s operating environment, and to provide added assurance for surrounding motorists, we separated the automated truck from other traffic on the road during the on-road trials. The truck’s lane was automatically and progressively closed to other traffic as the truck travelled along the road and re-opened behind the truck as it drove itself along the road.
PATH TO ON-ROAD TRIAL
Our trial was subject to an ADS trial permit, supported by a safety management plan and evaluation vehicle registration, and required National Heavy Vehicle Regulator standards exemptions and road access permits. The truck was trialled at the Australian Automotive Research Centre’s proving ground prior to on-road operations.
Safety drivers, who monitored the truck’s performance and would take control of the vehicle in line with safety protocols as required, were trained in the truck’s operations during the proving ground trials. The three safety drivers were all highly experienced heavy vehicle drivers.
ON-ROAD TRIAL OPERATIONS
We started our overnight trials by proving the truck’s core capabilities. We then added more complex manoeuvres and route extensions. Throughout, we were also testing how the ADS understood and responded to the real-time road-infrastructure data we provided.
RESULTS AND LEARNINGS
Our trial tested the performance of the automated truck and its ADS, as well as the dynamic dedicated lanes’ performance and other road-user’s compliance.
KEY RESULTS
- The ADS successfully navigated the full trial route and repeatedly demonstrated the ability to negotiate lane changes among other drivers.
- When ADS disengagements occurred, they were brief – less than 30 seconds on average – and they did not impede our ability to accomplish our trial goals.
- Data collected during the trial suggests the dynamic dedicated lane reliably delivered a simplified operating environment for the automated truck, with good compliance by other road users.
AUTOMATED TRUCK AND ADS PERFORMANCE
The automated truck used in this trial performed extremely well: 92% of the truck’s on-road test-environment travel was driven in automated mode. Further, the automated truck consistently performed as expected and the ADS safely handled the specific driving tasks we gave it. No safety incidents occurred at any stage of the trial. Data captured across the seven trials showed the ADS successfully navigated multiple operating environments, including:
- travelling in a dedicated lane alongside live traffic on a public motorway
- driving the full extent of both the Domain and Burnley tunnels (part of CityLink)
- performing lane changes in mixed traffic conditions, including multiple instances where the automated truck waited for other road users to pass before merging.
The ADS needed to be calibrated for the specific road conditions it encountered during the trial (for example, line-marking lengths and road-surface conditions) as these varied from the conditions the ADS had been programmed to handle. In CityLink’s tunnels, adjustments to the ADS’s method for understanding its current location were necessary due to the tunnel’s homogenous environment (that is, its smooth walls and the absence of regularly spaced landmarks).
TRIAL TECHNOLOGY PERFORMANCE
The dynamic dedicated lane was effective in both delineating the environment around the trial vehicle and reducing lane closure impacts for other road users. The dedicated lane greatly reduced the amount of interaction between the trial vehicle and other road users, and road-user compliance was excellent.
KEY LEARNINGS
- When operating on a new corridor, road-surface conditions can impede ADS performance. For optimal performance, sensors must be calibrated to accommodate applicable road conditions.
- Testing in environments representative of on-road conditions is essential prior to on-road deployment. Some improvements to proving ground capabilities would be beneficial in fully supporting I2V testing.
- Safety drivers played an important role in this early deployment, including by providing ADS performance feedback throughout the validation testing and on-road trials.
The trial found I2V data did enhance ADS performance. For example, the trial’s ADS was not trained to read LED signs. We used a road-infrastructure-to-vehicle technology solution to share real-time speed and lane status data with the ADS and this proved effective. We also found sharing real-time information on traffic ahead enabled the truck to ‘see’ around bends in the road and respond appropriately to traffic conditions.
PLANNINGS AND DELIVERY LEARNINGS
Our trial planning and delivery captured data relevant to vehicle manufacturers, road designers, operators, and regulators. Our learnings will help support overall advancement of automated trucks and road infrastructure technology and contribute to preparations for their arrival on Australian roads.
Our trial preparations included working with regulators and permit authorities to overcome challenges. For example, the positioning of the truck’s front-mounted sensors meant the truck was classified as overwidth. Installing overwidth signage on the truck in line with regulations would obstruct the sensor’s range and we needed to work with regulators to overcome this challenge. Early engagement between developers of automated vehicle technology and government regulators will be beneficial in finding solutions for these emerging challenges.
LOOKING AHEAD
Our trial results suggest well-instrumented managed motorways with mid-to-high freight volumes will be ideal candidates for early automated truck deployments. These major arterials have enough freight demand to attract automated truck operators and enough instrumentation to generate infrastructure data to support the automated trucks.
These findings are now helping Transurban, and our government and industry partners, better understand how roads and roadside technology can be leveraged to support automated trucks joining the mix of vehicles already moving freight on our roads.
Detailed results
You can learn more about our Automated Truck Trial and download the report here:
Trials | Transurban CAV trials
Transurban is a Centre for Connected and Automated Transport founding member.