 |
Ant-inspired robot called BionicAnt [1]
|
The life we see on Earth today has been molded by billions of years of evolution. Through random chance and lots of trial-and-error, organisms have evolved adaptations like fur, wings, and toxins to face the challenges around them. Biomimicry is a method that takes advantage of these adaptations by modeling innovations after nature. For example, Velcro was inspired by the stickiness of gecko feet and Japan’s high-speed bullet trains were modelled after kingfisher beaks.
Scientists are constantly looking for inspiration from nature, and many of them have found the biggest inspirations in some of the smallest creatures: ants. Even though individual ants are small and fairly simple, they can work together to accomplish great things like building complex underground tunnels, making life rafts during floods, carrying prey 100 times their own weight, and more. They are also able to make complicated collective decisions on where to build their new nests and how to transport large meals. Many of these decisions employ the use of pheromones: chemicals that ants can leave behind to indicate where they’ve traveled or explored. As pheromones accumulate in a certain nest or on a specific trail, there is a positive feedback loop that recruits more ants to explore the resource. As a result, local movements made by individual ants can lead to colony-wide changes and collective decisions.
By studying these simple behaviors in ants, mathematicians and computer scientists have come up with “ant algorithms” to solve more complex real-world problems--from managing road traffic to routing Internet network packages. Many of these algorithms mimic the feedback loops found in ant pheromone trails to make decisions based on local rules, without centralized control. Typically, these algorithms are then tested with computer simulations to see how reliable they are. For instance, scientists have tried to develop algorithms for sorting and aggregating objects modeled after the behavior seen in ant species like Pheidole pallidula, which gather their dead ants together in a heap, and in species like Leptothorax unifasciatius, which arrange their larvae according to their size.
Through simulations of artificial ants, the scientists were able to confirm that their algorithms worked, showing that they could be applied to other things like programming robots or analyzing data.
Scientists can also use simulations to understand ant behavior so that the behavior can be properly applied to solve bigger problems through biomimicry. A research team at Georgia Tech noticed that when fire ants dig underground nests, only 30 percent of the ants in a colony performed 70 percent of the work. To understand why the ants divide up the work so unequally, they had two different simulations of ants digging tunnels: in one simulation, the artificial ants divvied up the work in the same proportions as the real ants; in the other, the ants divvied up the work equally. Surprisingly, ants in the first simulation were able to create tunnels that were three times longer than those in the second simulation. It turns out that having only a few workers doing the heavy lifting at a time is more efficient in confined workspaces. The research team then applied its findings to robots that had to work along a narrow path, and the robots indeed performed better, with less traffic jams, when the workload was unevenly divided.
By working with real, simulated, and robotic ants, scientists can build on what they know and fine-tune their knowledge to tackle larger real-world problems through biomimicry. Scientists can change parameters in a simulation based on real ant observations to make the artificial ants more biologically realistic. Or they can use simulations to model scenarios that would be experimentally unreasonable, such as the nest construction of meat ant colonies, which normally take several decades. They can also give artificial ants extra bells and whistles that real ants don’t have, allowing them to perform even better than real ants. Robotic ants can, for example, be given long-term memory storage or be programmed with greater sensing capabilities through computer vision. The options are limitless when designs from nature are combined with tools from technology through biomimicry. By making connections between natural and man-made designs, biomimicry allows us to create biologically-inspired solutions while improving our understanding of the biological underpinnings. This interchange has led to innovations that seem to come straight out of science fiction, like tiny ant-inspired robots that are able to move a car weighing 3,900 pounds or life-like BionicANTS that can detect objects and communicate with each other through radio signals. Thus, by seeking to understand the biological world around us, we can take advantage of the wealth of knowledge provided by nature. Combining this knowledge with the human ability for creativity and innovation can then lead to incredible technological advances.
Further Reading
Christensen, D. L., Suresh, S. A., Hahm, K., & Cutkosky, M. R. (2016). Let's All Pull Together: Principles for Sharing Large Loads in Microrobot Teams. IEEE Robotics and Automation Letters, 1(2), 1089-1096.
Dorigo, M., Caro, G. D., & Gambardella, L. M. (1999). Ant algorithms for discrete optimization. Artificial life, 5(2), 137-172.
Handl, J., Knowles, J., & Dorigo, M. (2006). Ant-based clustering and topographic mapping. Artificial life, 12(1), 35-62.
Kammoun, H. M., Kallel, I., Alimi, A. M., & Casillas, J. (2011, April). Improvement of the road traffic management by an ant-hierarchical fuzzy system. In Computational Intelligence in Vehicles and Transportation Systems (CIVTS), 2011 IEEE Symposium on (pp. 38-45). IEEE.
Merkle, D., & Middendorf, M. (2002). Modeling the dynamics of ant colony optimization. Evolutionary Computation, 10(3), 235-262.
Zhao, D., Luo, L., & Zhang, K. (2009, October). An improved ant colony optimization for communication network routing problem. In Bio-Inspired Computing, 2009. BIC-TA'09. Fourth International Conference on (pp. 1-4). IEEE.
Media Credits
[1] Photo by Festo https://www.festo.com/group/en/cms/10157.htm
Scientists are constantly looking for inspiration from nature, and many of them have found the biggest inspirations in some of the smallest creatures: ants. Even though individual ants are small and fairly simple, they can work together to accomplish great things like building complex underground tunnels, making life rafts during floods, carrying prey 100 times their own weight, and more. They are also able to make complicated collective decisions on where to build their new nests and how to transport large meals. Many of these decisions employ the use of pheromones: chemicals that ants can leave behind to indicate where they’ve traveled or explored. As pheromones accumulate in a certain nest or on a specific trail, there is a positive feedback loop that recruits more ants to explore the resource. As a result, local movements made by individual ants can lead to colony-wide changes and collective decisions.
 |
Fire ants form life rafts during Hurricane Harvey [2]
|
Through simulations of artificial ants, the scientists were able to confirm that their algorithms worked, showing that they could be applied to other things like programming robots or analyzing data.
Scientists can also use simulations to understand ant behavior so that the behavior can be properly applied to solve bigger problems through biomimicry. A research team at Georgia Tech noticed that when fire ants dig underground nests, only 30 percent of the ants in a colony performed 70 percent of the work. To understand why the ants divide up the work so unequally, they had two different simulations of ants digging tunnels: in one simulation, the artificial ants divvied up the work in the same proportions as the real ants; in the other, the ants divvied up the work equally. Surprisingly, ants in the first simulation were able to create tunnels that were three times longer than those in the second simulation. It turns out that having only a few workers doing the heavy lifting at a time is more efficient in confined workspaces. The research team then applied its findings to robots that had to work along a narrow path, and the robots indeed performed better, with less traffic jams, when the workload was unevenly divided.
Fire ants inspire traffic-conscious robots [3]
Further Reading
Christensen, D. L., Suresh, S. A., Hahm, K., & Cutkosky, M. R. (2016). Let's All Pull Together: Principles for Sharing Large Loads in Microrobot Teams. IEEE Robotics and Automation Letters, 1(2), 1089-1096.
Dorigo, M., Caro, G. D., & Gambardella, L. M. (1999). Ant algorithms for discrete optimization. Artificial life, 5(2), 137-172.
Handl, J., Knowles, J., & Dorigo, M. (2006). Ant-based clustering and topographic mapping. Artificial life, 12(1), 35-62.
Kammoun, H. M., Kallel, I., Alimi, A. M., & Casillas, J. (2011, April). Improvement of the road traffic management by an ant-hierarchical fuzzy system. In Computational Intelligence in Vehicles and Transportation Systems (CIVTS), 2011 IEEE Symposium on (pp. 38-45). IEEE.
Merkle, D., & Middendorf, M. (2002). Modeling the dynamics of ant colony optimization. Evolutionary Computation, 10(3), 235-262.
Zhao, D., Luo, L., & Zhang, K. (2009, October). An improved ant colony optimization for communication network routing problem. In Bio-Inspired Computing, 2009. BIC-TA'09. Fourth International Conference on (pp. 1-4). IEEE.
Media Credits
[1] Photo by Festo https://www.festo.com/group/en/cms/10157.htm
[2] Photo by Tim Nowak https://www.popsci.com/fire-ants-take-advantage-hurricane-harvey-floods
[3] Video by Science News https://www.youtube.com/watch?v=cHo-ZJdxH_c&feature=youtu.be
No comments:
Post a Comment