In the future, we may have autonomous robots that follow scents to track down gas leaks, rescue disaster victims trapped in debris, or perform other duties. While the algorithms that drive such robots could perhaps just be made up from scratch, scientists from the University of Tokyo are instead looking to the insect world for inspiration. To that end, they recently created a two-wheeled robot that was successfully driven by female-seeking male silkmoths.

The research team chose the adult male silkmoth because it has a very specific mating dance that it performs in response to the detection of sex pheromones emitted by females. As it walks toward the female, it varies between straight and zig-zagged walking, taking a number of turns in the process, followed by its turning in a loop of over 360 degrees.

“The simple and robust odor tracking behavior of the silkmoth allows us to analyze its neural mechanisms from the level of a single neuron to the moth’s overall behavior,” explained Dr. Noriyasu Ando, co-author of a paper on the research. “By creating an ‘artificial brain’ based on the knowledge of the silkmoth’s individual neurons and tracking behavior, we hope to implement it into a mobile robot that will be equal to the insect-controlled robot developed in this study.”

The tests took place in a 180-centimeter (70.9-inch)-long wind tunnel, with a single-puff pheromone source at one end, and a robot-driving moth at the other, downwind from it. A couple of small fans mounted on the front of the robot helped channel the pheromones up toward the moth.

On board the robot, the moth was suspended over an air-floated polystyrene ball. As the insect responded to the pheromones, its legs instinctively moved in a fashion that would ordinarily allow it to walk across the ground toward the source. Given the situation, however, its legs instead spun the ball. An optical mouse sensor detected those ball movements, and used them to steer the robot along the same course that the moth wished to walk. A total of 14 male moths were put “behind the wheel” (or above it, to be precise), and they were all able to guide the robot to the pheromone source.

Interestingly, even when the robot was altered so that it pulled to one side, most of the moths were able to compensate to keep it on course. The scientists also introduced varying lengths of time delays between the turning of the tracking ball and the corresponding turning of the robot. They did so to mimic the delay between sensory input and motor responses that would inevitably be part of a purely man-made system. In particular, they were interested in exploring the relationship between the length of delay and the moth’s ability to reach the pheromone source – not surprisingly, they found that the greater the delay, the lower the success rate.

The research team’s paper was published today in the journal Bioinspiration and Biomimetics, and can be accessed online. Some of the robot trials can be seen in the video below.

View gallery - 4 images