Obsessed with technology?
Subscribe to the latest tech news as well as exciting promotions from us and our partners!
By subscribing, you indicate that you have read & understood the SPH's Privacy Policy and PDPA Statement.
News
News Categories

A*STAR researchers teach robot to navigate with simulated brain cells

By Koh Wanzi - on 22 Oct 2015, 11:12am

A*STAR researchers teach robot to navigate with simulated brain cells

Image Source: Pixabay

Researchers at the Agency for Science, Technology and Research (A*STAR) said that they have taught a robot to find its way around a room by simulating a cluster of virtual neural cells.

Humans and other animals navigate with the help of “grid” and “place” cells. Grid cells were discovered only in 2005, and as their name suggests, they help provide a detailed sense of position in a three-dimensional space. On the other hand, scientists have known about place cells for a far longer time – they were discovered in the 1970s by John O’Keefe, a Nobel-winning neuroscientist.

Place cells fire whenever an animal passes a familiar spot in its environment, while grid cells activate when it arrives at a location on a triangular grid of points. The researchers were thus able to track the functioning of the virtual cells by checking on their activation patterns.

The simulated cells were wired into a rough neural network for a wheeled robot, which was then given free reign of a 35 square meter office space. As the robot moved around, the team found that its artificial neurons fired as they should.

And even as researchers work to create smarter machines by mimicking actual neural networks, this is yet another instance of how machines could take a leaf from biology in order to become more intelligent.

However, the system created by the A*STAR researchers is far from perfect and it still has a long way to go compared to even conventional machine mapping methods. Nevertheless, the team believes that it is a potential first step in equipping future robots with adaptive navigation abilities. One potential application could be in self-driving cars, which stand to benefit immensely from the ability to adapt on-the-fly to changing road conditions.

Source: Technology Review via Engadget

Loading...