We've told you several times before about robotic bugs. We've also told you about robotic ants that don't look like insects but behave like them to help tell researchers how to design more efficient transportation networks. Now we've found some swarms of robot termites, a robot modeled after honeybees, and studies of ants done to produce stronger robot joints and muscles.
Inspired by termites' building skills, engineers and computer scientists at Harvard's Wyss Institute for Biologically Inspired Engineering, and its School of Engineering and Applied Sciences (SEAS), have created a crew of construction robots. The autonomous TERMES bots create 3D structures by modifying their environments, just as real termites move dirt around while building tall vents for their underground nests. Researchers presented results of the four-year study at the American Association for the Advancement of Science 2014 annual meeting and also published them in an article (purchase or subscription) in Science. (You can access this article for free via a link on the TERMES project page.)
Click on the image below to start the slideshow.
Termite mounds were the inspiration for Harvard's autonomous swarm of TERMES robots, which autonomously build towers, castles, and pyramids out of foam bricks. In the future, similar robots could lay sandbags in advance of a flood or perform simple construction tasks on Mars.
(Source: Eliza Grinnell, Harvard School of Engineering and Applied Sciences)
To simplify the processing load on the robots' onboard processors, and make the system more robust, robots and the foam bricks they carry were designed together. The bots operate using scalable, distributed artificial intelligence in a decentralized, multi-agent system. We've discussed swarming insect behavior several times as it applies to robots, and we've discussed centralized versus decentralized robotic systems in the context of modular, self-assembling, and self-reconfigurable robots.
The TERMES bots are aimed eventually at projects like working in disaster areas or extraterrestrial environments. For now, they carry foam bricks wherever they're needed for making a structure, or to build staircases that let the bots climb up to add more bricks to the structure. You can watch videos of the robots in action here, here, and here.
Autonomous robots modeled after a honeybee's simple nervous system, with a similar sensorimotor network, are being developed by German researchers at the Bernstein Fokus Neuronal Basis of Learning, the Bernstein Center Berlin, and the Freie Universitat Berlin. They've created a robot that navigates new environments by reacting to stimuli in its environment. Data comes in via a camera mounted on a tiny vehicle resembling a tank, and images are processed on-chip. Information is then passed to a spiking neural network via WiFi. The network controls the motors in the robotic tank's wheels to determine its motion and direction.
The robot operates via rules-based programming, for example, learning to approach some colored objects and avoid others. You can watch a video of the robot in action here. The researchers delivered a paper detailing their results at the 6th International IEEE/EMBS Conference on Neural Engineering (NER) last November.
How ants lift weights much bigger than they are is the subject of a study by Ohio State University researchers. They found that a common American field ant's neck joint withstands pressures of up to 5,000 times the weight of the ant's body. That's a lot more than previous researchers have guessed, which was more in the range of 100 times the insect's weight.
The engineers combined mechanical testing with computed tomography (CT), scanning electron microscopy (SEM), and computational modeling to get a grip on the relationship between the ant neck joint's structure and its functioning. Then they developed a 3D model of the neck joint's structure, including its components. Using finite element analysis (FEA), they discovered that the neck joint's critical failure point is the transition between neck and head where the hard exoskeleton meets a soft membrane. The team published its results in an article (purchase or subscription) in the Journal of Biomechanics.
Similar artificial joints combining hard and soft parts might be designed to make future robotic ants that can lift big weights, either on earth or in space. Larger robots based on a similar design wouldn't work on Earth, since weight increases much faster than muscle strength as volume increases. But they might be practical in the microgravity of space. The team also plans to study the antís muscles, possibly using magnetic resonance imaging (MRI). The members want perform simulations to figure out how to scale up similar robotic structures.