Dragonfly vision used in driverless cars could predict traffic movement

Departments - Rearview


Researchers from the University of Adelaide in South Australia and Lund University in Sweden have found a neuron in dragonfly brains that anticipates movement. The properties of the target-detecting neurons are being replicated in a small robot in Adelaide to test its potential for artificial vision systems used in driverless cars.

Research supervisor and lecturer at the University of Adelaide’s Medical School Steven Wiederman says the new discovery would have an impact on driverless cars and other robotic vision systems.

“It is one thing for artificial systems to be able to see moving targets, but tracing movement so it can move out of the way of those things is an important aspect to self-steering vehicles,” Wiederman says. “What we found was the neuron in dragonflies not only predicted where a target would reappear, it also traced movement from one eye to the other – even across the brain hemispheres.

The research team, led by University of Adelaide Ph.D. student Joseph Fabian, found that target-detecting neurons increased dragonfly responses in a small focus area in front of a moving object’s location.

If the object then disappeared from the field of vision, the focus spread forward over time, allowing the brain to predict where the target would reappear.

The neuronal prediction was based on the previous path the prey had flown.

Wiederman explains how this phenomenon was not only evident when dragonflies hunted small prey but when they chased after a mate as well.

This is similar to when humans judge the trajectory of a ball as it is thrown to them, even when it is moving against the backdrop of a cheering crowd.

This is the first time a target-tracking model inspired by insect neurophysiology has been implemented on an autonomous robot and tested under real-world conditions.

The University of Adelaide's autonomous robot testing its sensing techniques derived from dragonflies.

South Australia has a history of involvement with autonomous car research and in 2015 held the first driverless car trials in the Southern Hemisphere.

It hosts several autonomous car companies including Cohda Wireless and its innovative V2X (Vehicle to everything) technology and RDM Group, which opened its Asia-Pacific headquarters in Adelaide earlier this year.

University of Adelaide Researcher Zahra Bagheri says there is growing interest in the use of robots for applications in industry, health and medical services, and entertainment products. “However, our robots are still far behind the accuracy, efficiency, and adaptability of the algorithms that exist in biological systems.”

“Nature provides a proof of concept that practical real-world solutions exist, and with millions of years of evolution behind them, these solutions are highly efficient,” Bagheri adds.

A study on the implementation of CSTMD1 into the robot was published in the Journal of Neural Engineering (goo.gl/8mrrHx).

University of Adelaide


Lund University


The research project is an international collaboration funded by the Swedish Research Council, the Australian Research Council (ARC), and STINT — the Swedish Foundation for International Cooperation in Research and Higher Education.