Donate to Science & Enterprise

S&E on Mastodon

S&E on LinkedIn

S&E on Flipboard

Please share Science & Enterprise

Robot Navigates New Spaces with Bat-Like Sonar

Southern bentwing bat

Southern bentwing bat (Steve Bourne, Wikimedia Commons)

7 September 2018. A terrestrial robot demonstrates an ability to map and move through new environments using an ultrasound system inspired by bats, as well as decision-making algorithms. The Robat, as the system is called, is described in yesterday’s issue of the journal PLOS Computational Biology.

The Robat, an invention of engineering graduate student Itamar Eliakim at Tel Aviv University in Israel, uses neurological and biometric capabilities of bats to provide self-contained navigation for autonomous robotic devices. Robots sent on search and rescue missions through terrain disrupted by rubble from natural disasters or combat, for example, would likely encounter many obstacles in an environment that could quickly change. The authors note that other autonomous robots are able to travel through newly-encountered territories, but with most of these devices each obstacle is navigated independently, while the Robat stores and maps information for later use.

The system emulates the echolocation abilities of bats that use sound waves for navigation. The Robat is built on a commercially-available Komodo robotic platform for indoor and outdoor monitoring on rough terrain, made by Robotican, an Israeli company. The off-the-shelf device has built-in sonar, as well as a camera, GPS, and Bluetooth and WiFi connectivity. Eliakim and colleagues devised a module for the Robat with a speaker to emit ultrasound waves and two antennae to record the echoes, in effect emulating a bat’s mouth and ears. Every 0.5 meters, the Robat stops and emits ultrasound waves in 3 directions, then collects the echoes.

The echoed signals are collected and analyzed in mapping software, with new objects integrated into the overall map of terrain. Objects encountered are classified with algorithms in a neural network, a form of artificial intelligence, developed by the researchers. The classifications help determine, for example, if an object is an obstacle that can be circumvented or a wall blocking further advances. The neural network was trained with more than 1,400 objects on the Tel Aviv University campus.

The team tested the Robat in two greenhouses in Tel Aviv University’s botanical garden, 60 and 200 square meters respectively, that the device had not previously encountered. The Robat traveled through both greenhouses, successfully navigating the terrain, and moving around obstacles put in its way. The maps of the greenhouses produced by the Robat generally conformed to contour maps made independently, although the researchers admit that bats’ natural navigation is more precise.

“Our Robat,” says Eliakim in a university statement, “is the first fully autonomous, bat-like biorobot that moves through a novel environment while mapping it solely based on echo information. This information delineates the borders of objects and the free paths between them.”

The Robat is a joint project of Tel Aviv University’s zoology and engineering departments. Eliakim worked in the labs of Yossi Yovel that researches the neuro-ecology of bats, and robotics and engineering professor Gabor Kosa. The following video demonstrates the Robat, showing simultaneous signal processing and mapping.

More from Science & Enterprise:

*     *     *

Comments are closed.