The yellow slime mould Physarum polycephalum has been used to control a six legged motorised robot by researchers at Japan’s Kobe University.
This type of mould can be found in damp, shady areas of woodland, where it exists as a giant cell of up to 20 metres in length. They feed on rotting leaves and plant matter and when exposed to white light can change their growth direction in avoidance. By avoiding the strong white light during daytime they are able to locate more suitable places to feed without becoming too dry.
The researchers made use of this photophobic attribute as a mode of control for their robot as it provided them with the context-sensitive mode of information processing they were looking for. Typically, computers would have been used for this task yet they can often be too cumbersome and inefficient for practical use. Biological systems however excel at interacting and responding to dynamic environments which makes them ideal for this kind of application.
The slime mould (top left) is grown on a six welled circuit (top right). Light exposure to wells on the circuit moves the robot’s legs (bottom image).
In humans and other animals information processing is via the nervous system which ensures an appropriate rapid response can be made to a particular stimulus. In P. polycephalum vein-like tubules which shuttle cytoplasm back forth perform a similar role (albeit much more slowly). The rhythm of these back and forth oscillations dictates the thickness of the tubules and the overall movement of the cell.
By growing the organism as a single cell in precisely shaped wells on an agar plate “slime circuits” could be made. These circuits possessed six channels in a star conformation, with each channel corresponding to a leg on the robot’s chassis. Light exposure detected by sensors on the robot’s body relayed signals to a computer prompting a light to be shone onto a particular channel. This initiated the movement of the robot’s leg as the slime evaded the white light.
Over time, when combined with the natural oscillations within the slime mould’s tubules a walking motion was observed which would cause the robot to evade the light as an incarnation of the slime mould’s responses.
Although the work carried out here does prove the concept that biological and mechanical systems can be integrated it is still a technology which is in its infancy. The outlook does however look promising due to the novel attributes of this style of design. Features such as the organism’s self-repair capability and decision making capacity are definitely appealing, and the low power requirement of the robot is also an advantage.
Work has already begun to combine the circuit and sensor in one unit, eliminating the need for the cumbersome computer that is currently necessary. Consequently, in the future we may begin to more of these amalgamations of microorganism and machine that could even be tailored for a particular task.