Robophysical study of excavation in confined environments

• Published in: Artificial life and robotics Vol. 21; no. 4; pp. 460 - 465
• Main Authors: Linevich, Vadim, Monaenkova, Daria, Goldman, Daniel I.
• Format: Journal Article
• Language: English
• Published: Tokyo Springer Japan 01.12.2016; Springer Nature B.V
• Subjects: Artificial Intelligence; Cellular automata; Computation by Abstract Devices; Computer Science; Confined environments; Confined spaces; Control; Excavation; Excavators; Insects; Mechatronics; Plugging; Robotics; Robots; Special Feature: Original Article; Tunnels (transportation); Ants; Autonomous excavators; Confined spaces; Robots
• ISSN: 1433-5298; 1614-7456
• DOI: 10.1007/s10015-016-0317-2

Soil-dwelling social insects build complex nests. Nest excavation is performed by multiple animals simultaneously and is governed by local interactions of the workers with other nest-mates and their surroundings. To investigate collective confined excavation challenges, we built groups of robotic excavators capable of performing hours of autonomous tunnel excavation in a model cohesive granular medium. Excavator behavior was governed by a simple set of rules triggered by interactions with the surrounding environment and other robots. The rate of tunnel growth and energetic costs of excavation were measured for groups of different numbers operating in wide and narrow tunnels. To extend the results to systems with large numbers of robots, we developed a cellular automata model. Experiments and simulations showed that in sufficiently wide tunnels an increase in the size of the excavating group increased the excavation rates without a significant increase in the energy consumption per robot. A decrease in the tunnel width resulted in a decrease in the excavation rates and increase in the energetic costs of excavation. We attribute this effect to the emergence of multiple time-consuming interactions (clogs) among excavating robots in the confined spaces. Although in all situations clogs were resolvable, clog resolution took longer in the systems with larger number of robots and narrower tunnels. We expect that our robotic system can be used to investigate the behavior of social insects in confined spaces as well as inspire more sophisticated search-and-rescue robotics.