A New Twist in Robotic Exploration

For many applications in robotics, a critical first step involves mapping the environment. After all, a robot is going to have a very tough time finding its way around, or completing any meaningful tasks, if it cannot first understand the world around it. Traditionally, this means collecting data from cameras, lidar, and any number of other sensors. The sensor measurements then need to be processed, most commonly by resource-intensive machine learning algorithms, to make sense of it.

The sensors, and the computing resources needed to process the collected measurements, can be quite expensive and require a large amount of energy for operation. Moreover, developing and refining the processing algorithms can be very labor-intensive. The cost and level of expertise required to develop such systems can leave them out of reach for many. Further, the bulk added to a robot by all of these components can make them unsuitable for certain applications, especially where weight or physical size must be restricted.

To democratize access to robots capable of mapping out complex environments, further technological advancements are needed. A step in that direction has recently been taken by a team of researchers at North Carolina State University. They have approached the problem from a completely different direction, eschewing sensors, advanced algorithms, and compute resources altogether. The result is a soft robot with no traditional control system, but rather with inbuilt physical intelligence that enables it to explore an unknown area and map out its boundaries.

Dubbed twisted ringbots, these little robots are made of a ribbon-like liquid crystal elastomer. The ribbon is twisted up, then joined together at one end to form a loop. Importantly, the elastomer is also temperature-sensitive. When placed on a warm surface, the material touching that surface contracts while the area exposed to the cooler air does not. This gives rise to a number of interesting effects that cause the robot to simultaneously roll forward, spin along its central axis, and travel along an orbital path around a central point.

When the robot runs into a boundary, like a wall, it will closely follow the path of that boundary. This factor could be very useful in mapping out environments. Since the robots move along a relatively simple path, that would limit them to mapping out simple environments. However, by modifying the material, like by twisting the elastomer in different ways, the robots can be made to follow different paths. In this way, a large number of these robots could map out even very complex environments by working together. And since the robots are inexpensive to produce, it would be practical to utilize many for such a purpose.

In a series of tests, the twisted ringbots were demonstrated as being capable of locating and following the boundaries of a variety of confined spaces. This bodes well for their future as mapping devices, however, they will not be applicable to every situation. The Achilles’ heel for these robots may be in their method of locomotion. In order to move, they must be placed on a surface that is at least 131 degrees Fahrenheit. Unless this factor can be adequately addressed, these robots will likely be relegated to a limited number of niche applications.