Scientists have found a way to let robots move through small gaps by attaching a shell reminiscent of that of a cockroach.
The cockroach is a hugely resilient animal, but also one of the world’s most hated. Now the insect has inspired a new design for a robot to negotiate difficult spaces, writes Robert Ferris for CNBC.
Nature inspires robots
Researchers have been trying to find ways of employing Nature’s ingenious systems in robots, mostly to avoid obstacles.
“The majority of robots deal with obstacles by avoiding them—often using sensors to map out the environment and heavy computation to plan a safe path to go around obstacles,” said lead researcher Chen Li. “This approach has been very successful (for example, Google’s self-driving car). However, it does have limitations. First, when the terrain becomes densely cluttered, a clear path cannot be planned because obstacles are just too close to each other.”
Researchers were studying cockroaches in order to find out how they were able to move in narrow spaces, and discovered that cockroaches are able to get through difficult terrain thanks to their bodyshape. The study of cockroach movement led to the design of a similar system which allows a robot to move with greater agility.
Study leads to improved performance for simple robots
First of all the team filmed cockroaches as they moved through a series of densely packed obstacles. The rounded body was found to act as a guide-rail, allowing the insect to move through narrow gaps without burning that much extra energy upon making contact with an object.
Next the researchers constructed a rounded shell and attached it to the top of a robot, and watched as it rolled onto its side and slid past obstacles in the same way as the insect. The test robot had no obstacle-sensing system, so the improvement in performance was all down to its shell.
Such designs could be used for robots which will perform search-and-rescue missions or work in precision agriculture. The shell would allow them to navigate thick vegetation or rubble without expensive extra sensors or complex software.
Detailed findings from the study can be found in the journal Bioinspiration & Biomimetics.
