Forget driving dogs - Japanese scientists have created a robot which can be controlled by moths.
Researchers from the University of Tokyo have developed a two-wheeled robot which a male silkmoth (Bombyx mori) can control, by walking on a rotating polystyrene ball similar to the trackball of a computer mouse.
The moths were tasked with steering the machine through a 1800 millimetre wind tunnel to a point where female moth sex pheromones were being excreted.
Two 40mm fans were attached to the wind tunnel to divert the pheromone towards the on-board moth, which the researchers believed would be comparable to how the wings of the silkmoth would generate air flow across its antennae.
In all, 14 moths were used, and each one was able to manoeuvre the robot to where the scent was coming from.
The male silkmoth was chosen as the driver of the robot because of its characteristic mating dance when reacting to the sex pheromone of the female. Once the male is stimulated by the pheromone it exhibits a distinctive walking pattern: straight-line and zigzagged walking consisting of several turns followed by a loop of more than 360 degrees.
The researchers also introduced a turning bias to the experiments, changing the power of one of the robot's two motors so it veered towards one side when moving. This put the silkmoth into an extraordinary situation and required it to adapt and change its behaviour.
Lead author of the research, Dr Noriyasu Ando, said the best way to study adaptive behaviours in insects is to put them into extraordinary situations.
"The turning bias in our study is analogous to a situation in which we try to ride unbalanced bicycles. We need training to ride such bicycles smoothly but the silkmoth overcomes the situation with only simple and fast sensory-motor feedbacks," Dr Ando said.
The researchers hoped the findings, which have been published in the journal Bioinspiration and Biomimicry, could lead to the creation of autonomous robots tasked with locating the source of chemical leaks and spills.
"Most chemical sensors, such as semiconductor sensors, have a slow recovery time and are not able to detect the temporal dynamics of odours as insects do," Dr Ando said.
"Our results will be an important indication for the selection of sensors and models when we apply the insect sensory-motor system to artificial systems."