Artificial intelligence for powering the application of spatial information

Artificial intelligence for powering the application of spatial information

251
0
SHARE

A future that readily facilitates automation of creating, sharing, curating, delivering and using knowledge to support the emerging digital economy and the rise of spatially-aware and equipped citizens

Machines, robots, computers. These are some of the words that come to mind when we think about artificial intelligence. We think of driverless cars and trucks, drones and new types of manufacturing processes and improved healthcare. At the cutting-edge of science and technology, artificial intelligence does spark debate amongst the community and researchers about the pathway towards technological singularity — a runaway growth that may result in unfathomable changes to civilization.

Amongst these debates however, investment is growing rapidly in the best brains to increasingly build intelligence and deep learning in the spatial infrastructures and systems that support economy-wide and society-wide needs. At the Australia and New Zealand CRC for Spatial Information (CRCSI) we build on these global capabilities using location-based technology to match the science and technology applications of location-enabled intelligence into areas that make — or will in the future make — better business sense (ease-of-doing-business) for our partners.

DriverlessTractor in Sugar Trials, December 2016
DriverlessTractor in Sugar Trials, December 2016

One of the core research themes at the CRCSI is the modernizing of Australia and New Zealand’s spatial infrastructure to support a real-time, precision, mobile world. Our future will see essential data (usually provided by government including cadastre, addressing, administrative boundaries, transport and land cover amongst others) delivered on-the-fly in a format that is usable by anyone — the community, private sector and governments. We see a future that readily facilitates automation of creating, sharing, curating, delivering and using knowledge (not just data or information) to support the emerging digital economy and the rise of spatially-aware and equipped citizens.

We envision the creation of a Spatial Knowledge Infrastructure (SKI) and will soon be releasing a discussion paper that is designed to stimulate debate about the essential features of an SKI.

The research of the CRCSI and our partners (government, universities and the private sector) is formulated to ensure Australia and New Zealand is building the capabilities to meet current and future challenges.

Positioning research

This research is seeking to underpin Australia’s National Positioning Infrastructure, led by Geoscience Australia, with real-time sub-decimetre accuracy anywhere outdoors across the continent. The challenge has been to solve the signal processing problems that will allow Australia to simultaneously use the best combination of signals from the six Global and Regional satellite systems that operate over Australia; GPS, QZSS (Japan), IRNSS (India), Beidou (China), Galileo (European Union) and GLONASS (Russian Federation).

For example, the Japanese QZSS satellite has performed well in two precision agricultural trials in Australia, the most recent in December 2016 in our sugar growing region in the tropics of northern Queensland, and the previous trial of January 2015 in the rice growing area of southern, cool temperate, Australia. The trials were conducted without the use of mobile phone connectivity. Both trials sought confirmation of the reliability of the QZSS LEX signal using the transmission of PPP-RTK corrections. The first trial saw an autonomous (driverless) tractor perform normal activities (driving, tilling, spraying) in a rice growing scenario controlled entirely through the communications channel of QZSS that operated alongside the positioning channel. The trial tested the proof of concept that the real-time communication between the QZSS and the tractor was stable and accurate to 5 cm.

Driverless image

The sugar growing trial sought to establish signal reliability and test accuracy from the QZSS satellite to the on-ground autonomous tractor in challenging local ionospheric and tropospheric conditions. The trial improved on the QZSS LEX signal availability and the transmitted PPP-RTK corrections compared to the earlier trials in 2015.

Underpinning the accuracy of precise positioning is the development of Australia and New Zealand’s first dynamic datum. Australia is literally on the move, with a recorded 1.8 metres to the north east in 20 years. Creating a dynamic datum — rather than the fixed GDA94 (Geodetic Datum of Australia) we currently use — will provide ongoing corrections to the continuing tectonic plate movement that creates inaccuracies for our positioning, thereby removing the effect of continental drift from our map base. Known as GDA2020, this new datum will provide sub-decimetre accuracy for industries such as surveying, construction, agriculture, environmental management, transport, insurance, emergency services, telecommunications and scientific research.

Early in 2017, the Australian Government announced a $12 million investment in an SBAS testbed (satellite based positioning augmentation system) and in February the New Zealand Government announced that it would be joining the trial with a $2M investment. Partnering with Geoscience Australia and LINZ (Land Information New Zealand) will be Lockheed Martin, Inmarsat and GMV. The CRCSI will draw on this global expertise to implement SBAS testing in Australia and New Zealand for the first time through a range of companies supported by a number of universities. This technology will provide Australia and New Zealand with a number of benefits — improved safety, productivity and efficiency gains — benefits that countries and regions like the US, Europe, China, Russia, India and Japan are now enjoying.

3D FAST analysis using spatial technologies in 3D imaging, deature extraction and mapping
3D FAST analysis using spatial technologies in 3D imaging, deature extraction and mapping

SBAS utilizes space-based and ground-based infrastructure to improve and augment the accuracy, integrity and availability of GNSS signals such as GPS. Initial industries to be involved in the Australian testing include road, rail, agriculture, aviation, construction, mining, maritime, spatial and utilities.

Rapid analytics

Harnessing machine-to-machine expertise to turn large amounts of information and data into fit-for-purpose knowledge forms part of our newest research program — Rapid Spatial Analytics. Working with partners that have access to large amounts of data (location, movement, and people) allows our research community to develop the models that become applicable in real-world uses. The research seeks to reduce the time to produce critical information products initially focusing on:

  • Developing a non-technical, automated process manage the workflow of mobile LiDAR data acquisition and to determine the quality of the data quality from mobile LiDAR
  • Building automated land valuations for Valuer General and governments and visualizing the results
  • Rapid data analysis of the movement of people in highly concentrated areas such as train stations
  • Moving workplace culture from open data to open analytics
  • Creating shareable and transparent open analytic workflows.

Agriculture

Our research delves into the production chain of the Australian red meat industry. Whilst there are currently numerous commercially available touchpoints along the value chain from livestock in the paddock to meat on a plate, this new line of research activity will embed spatial intelligence along the entire supply and value chain. The research is implementing a whole lifecycle of analyzed data that provide new opportunities to producers, processors, retailers and exporters to increase quality per kilogram for higher return on investment along the chain. The generic thinking being developed here is capable of application in many critical food chains including grain, horticulture and sugar.

Health

The health research is focussed on novel modelling, tools and capabilities to improve decision support, business intelligence and resource management for the health sector. The Health Program underpins its approach to using spatial information by answering the question, “does my location affect my health?”

Key outcomes include; the creation of the Atlas of Cancer in Queensland which resulted in public policy reforms, including a landmark increase in Queensland’s Patient Travel Subsidy Scheme in 2013 and now led to the National Cancer Council supporting the establishment of a program to create a National Cancer Atlas, with both Atlas’ underpinned by Bayesian statistical analyses; a data reporting and visualization tool, Healthtracks, that services hundreds of users of health data in the Western Australian public health system; and 3DFAST, a clinical facial analysis tool to map genetic diseases.

Intelligence is integral to the future of spatial research and the combination will help build capacity within nations that will better service the needs of society.