HD mapping Australia’s CAV future: Final report

The most critical contributor to the improved performance of automated vehicles is the availability and use of prior maps of the environment. iMOVE, along with Queensland Department of Transport and Main Roads, QUT, and RACQ, ran a project on how best to push the creation, monitoring and maintenance of HD maps to facilitate automated driving, and have published a final report on the findings.

The project was entitled HD maps for automated driving – literature review, with research led by QUT’s Professor Michael Milford. This project followed on from an already completed project, How automated vehicles will interact with road infrastructure. Both these projects investigate what it will take to have connected and autonomous vehicles operating efficiently and safely in Australia.

Why do we need HD maps, and what do they do?

Prior maps encode road‐level features such as the position of street signs and lane markings, with up‐to centimetre accuracy. This allows for the vehicle to verify the data received by its sensors against the prior map, and to even “fill in the gaps” where the incoming sensor data is limited due to difficulties such as rain or occlusions.

Not only does this level of detail, reliability, and accuracy likely need to exist before the vehicles are on our roads, the work to have them remain so is ongoing. In this we will need collaboration between car manufacturers, map makers and Governments, ‘… with agreements concerning geospatial data storage and sharing critical to the future success of automated vehicles.’

Key questions and investigations

The scope of this project was defined by the following questions:

• What is being done globally to investigate the use of prior maps in automated driving?
• What are the EU, US and other governments assuming?
• What are the likely industry models, considering the following specifications of:
  • Type of content
  • Extent
  • File formats
  • Initial acquisition approaches
  • Maintenance requirements
  • Limitations
  • Commercial models
• What models are being implemented or considered for the [Australian] government’s role in prior maps, which could include being the creator, curator or consumer of maps?
• What is being assumed of the government?

What’s next?

As was noted in our article published when this project began, ‘The outcome of this project is expected to inform the scope of further study on the matter.’

In the final report, two avenues for future work are provided:

1. Sovereign HD Mapping Capability

An automated vehicle trial, for example using the Zoe 1 and 2 research car platforms, could evaluate the capability of an automated vehicle to safely navigate Queensland roads using a continuously available prior HD map.

2. Collaborative HD Mapping

Investigation of the utility of industry‐government collaboration on HD mapping. Again, an on-road trial using the Zoe vehicles, but rather than building the maps in-house, instead working with industry to trial use of their state‐of‐the‐art HD map solutions.

These are but two ideas, ‘intended to act as exemplars for what could be done to further studies and understanding in this area, but there are many other opportunities beyond just these two.’

Download the report

We’ve made the final report, Map creation, monitoring and maintenance for automated driving available for download. Click the button below to download the document.

Also available for download is a HD Maps Fact Sheet.