Researchers in Singapore, led by Haizhou Li, a professor at the Agency for Science, Technology and Research (A*STAR) simulated two cell types that the brain uses for navigation, called “place” and “grid” cells, by creating a simple 2-D cell model in software. They used the cells to enable a small-wheeled robot to navigate.
Artificial grid cells may provide an adaptive and robust mapping and navigation system, mimicking more complex activity in the brain. Artificial neural networks, inspired by the brain, are gaining in popularity to train robots to perform such tasks as object recognition and grasping.
Place cells and grid cells activate when an animal passes the same spot or arrives at any location on a triangular grid of points. The cells are believed to provide animals with a sense of the world around them and of their location. The discovery of these cells earned three scientists the Nobel Prize in Medicine in 2014.
Researchers tested the approach on a robot in a 35-square-meter office space. The robot roamed around the space and its artificial place and grid cells functioned in a comparable way to their biological counterparts. The navigation system could offer advantages over conventional systems.
Question or comments on this article? Contact an editor: engineering360editors@ihs.com
