Re-Evolving Nature’s Best Positioning Systems for People and Their Machines

Summary

This project is an Australian Research Council Laureate Fellowship, awarded to Professor Michael Milford in 2021. Laureates are the premier fellowship scheme in Australia, and a maximum of 17 are awarded in any year. They support the fellow and large research teams for 5 years. In this case, it’s a $2,716,000 fellowship (plus approximately $680,000 cash funding from QUT)  supporting 4 postdocs, 8 PhDs, 25 honours students and others – as well as other vital research support including robotic equipment. The focus is on basic and applied research in strategic focus areas for the Australian Government, growing local and international collaboration, attraction and retention of top research talent, and training and mentoring of early career researchers.

Detailed Description

“to find where you are going, you must know where you are.”

John Steinbeck (1980). Travels with Charley in Search of America

Everything that moves is defined and limited by its ability to navigate the world in which it exists. Knowing where you are located in the world is a key navigational capability for people, animals, and both autonomous and human-operated platforms ranging from self-driving cars to aircraft. But accurate and trustworthy positional knowledge has widespread potential implications beyond navigation: it can, for example, allow life-and-death decisions in defence and in tracking the spread of global pandemics. Both the potential of and problems like privacy posed by modern social media networks are due in significant part to how positional tracking information is generated, used, and shared. With better positioning information, robotics, autonomous vehicle, and AI technologies of the future could transform and improve the competitiveness of current strategic industries like mining and agriculture. It could bring new capabilities in manufacturing and medical technologies, and re-invigorate sectors under pressure like aged care and transport.

Current positioning technologies are fundamentally incapable of meeting current and future positioning demands because they do not and will never meet the required performance, reliability and trustworthiness targets. By re-evolving the best positioning systems developed in nature to incorporate modern technologies and fulfill modern human requirements, we will create a suite of next-generation positioning technologies that address a sovereign risk and keep Australia’s industries competitive.

The overall aim of the Fellowship is to build a new generation of positioning technologies that will give Australia two crucial capabilities:

  1. A sovereign and uninterruptable positioning technology suite that gives Australian defence forces, industry, government, and society the positioning capabilities offered by current paradigms like Global Positioning Systems (GPSs) without risky dependence on satellites, and
  2. A new utility-maximisation approach to positioning technology development that makes these systems fit-for-purpose, interpretable, privacy-preserving, and trusted by their end-users, among them people, social networks, robots, drones, and autonomous vehicles.

Developing these two capabilities will both address a growing strategic risk for Australia and create a suite of new industrial and societal opportunities for Australia.

Research Programs

Research Program 1: Re-evolving Nature’s Best Positioning Systems

The animal kingdom hosts an astounding range of positioning capabilities, but no single organism reigns supreme. We will develop hybrid models of the best-performing and best-understood neural and behavioural positioning components from multiple species, including humans, other mammals, and insects.

Research Program 2: Trusted and Explainable Positioning through Auditable Neural Networks

Current positioning solutions based on machine learning cannot achieve enough performance while retaining sufficient provability and explainability characteristics for their transformative use in industry and society. A new class of ultra-compact, high-performance biologically inspired neural network approaches can potentially provide the required level of performance and a sufficiently compact network structure to be interrogated and audited for validation and verification.

Research Program 3: An Enriched and Functionally Relevant Understanding of Position and Place

The knowledge of where a human or robot is located moves far beyond the sparse co-ordinate representation provided by GPS or an INS system. How that entity navigates and makes decisions depends on a richer understanding of the place it finds itself in – its characteristics, dynamics, and functional relevance to the tasks being performed.

Research Program 4: Privacy-compliant and Cyber-secure Positioning

Creating a new generation of positioning technology that meets technical performance requirements is not enough: the utility of these systems will critically depend on whether that positioning information can be shared whilst maintaining the privacy and security of that information. Rather than attempt to retroactively deal with these problems in a separate privacy or cybersecurity framework, we’ll develop the architecture of the positioning systems themselves to achieve these goals.

Media and Links

Opportunities

The Laureate Fellowship will build a large, diverse research team.

We are eager to talk to those interested in any of the following positions:

  • 4 Research Fellows
  • 8 PhD students
  • 25 final year honours students

Funding / Grants

  • ARC Laureate

Partners