The Development of a Framework and Algorithms for a Distributed Heterogeneous Robot Team
he value of a group of entities collaborating as a However, small robots are more effective in other domains. team has been proven many times in many do- Small robots can potentially crawl through pipes, access collapsed mains. In the military, a team of men with spe- buildings, and hide ininconspicuous spaces. This increased accescialized skills and limited resources coordinate to sibility dramatically impacts the effectiveness of the team in some produce a combat unit with incredible force and surveillance and exploration tasks. However, with small size corange. Nature abounds with examples of predators who coor- mes the disadvantages of limited mobility range, limited energy dinateas a team to hunt prey that are stronger and swifter. availability, and possibly reduced sensing, communication, and These examples illustrate that the coordination of individuals computation ability due to size and power constraints. Because with similar or disparate abilities can produce a team with abil- limitations in size are immediately extended to power and processing capabilities, it wasrealized early that our robots would ities greater than the sum of its parts. The same advantages can be extended to a group of robots have to coordinate to achieve any useful tasks. The main focus of our research is coordinating as a team to share into develop a framework and algoformation and resources. In such a By LUIS E. NAVARRO-SERMENT, rithms for a group of small robots that team, a task isnot completed by a ROBERT GRABOWSKI, can effectively achieve the functionsingle robot but by the team of colCHRISTIAAN J.J. PAREDIS, and ality of a larger robot while retaining laborating robots. Team members PRADEEP K. KHOSLA the ability to operate in new doexchange sensor information, colmains. This article describes the delaborate to track and identify targets, or even assist each other to scaleobstacles. As for sensing, sign and construction of a team of 7 × 7 × 7-cm robots called by coordinating its members a team can exploit information “millibots” (Fig. 1). We show how the team can exploit collaboderived from multiple disparate viewpoints. Even a single ro- ration to perform missions such as mapping, exploration, surveilbot, though equipped with a large array of different sensinglance, and eventually support rescue operations. modalities, is limited at any one time to a single viewpoint. Moreover, a team of robots can simultaneously collect infor- The Millibot Team mation from multiple locations. There are many tasks for To attain reasonable functionality for robots on a small scale which distributed viewpoints can be exploited, such as sur- requires careful attention tothe development of the architecveillance, reconnaissance, and rescue. ture of the team. As size is reduced, it becomes increasingly One factor that determines team capability is the physical size difficult to outfit a single robot with all the necessary sensors of its members. A team composed by large-sized all-terrain vehi- required to complete the mission. Furthermore, limitations in cles (ATVs)would do very poorly mapping the inside of a build- power and size further reduce the range and capabilities of ing, whereas the same team would excel when exploring and these sensors. Our approach to overcoming the disadvantages surveying urban areas. Conversely, small robots—even in teams— imposed by small robot size is to exploit the properties of speare inappropriate in large, open spaces, suchas fields or forests. cialization and collaboration.
IEEE Robotics & Automation Magazine
Specialization is necessary to allow the team as a whole to optimize on size and resources—specifically power. For a robot to be mobile, it must carry its own power source. Considering that the battery volume for robots on the centimeter scale...
Leer documento completo
Regístrate para leer el documento completo.