Citation Context

...aintenance controllers for greater robustness. The robots jointly switch between different controllers, so as to address intermittent failures in sensing or communications. The use of teamwork software in robotics has been reported before [6], but its use for integration of different control schemes is novel. 2. RELATED WORK In formation-maintenance tasks, the objective is to move a group of robots on a desired path, while they maintain their relative position with respect to their peers, according to a desired geometric shape. Many formation maintenance methods have been investigated. (e.g., [2, 8, 5, 7, 3]). These vary in their operating assumptions, their reliance on specific sensors, and their ability to handle different types of faults. Some researchers have examined integration of multiple methods. Fierro et al. [4] analyzed the stability of different controllers, and proposed using manually-constructed control targets to allow up to three robots to switch between alternative schemes, without relying on communications. Fredslund and Mataric [5] describe a method that supplements SBC control by communications, for robustness, thus fusing SBC and communications-based control. Elmaliach and K...

Citation Context

...intaining relative positions with respect to each other (typically describing a specific geometric shape). Indeed, there exists vast literature on various techniques for maintaining formations in a variety of settings. Different controllers have advantages and disadvantages which can be complementary. Little attention has been given to the possibility of integrating together multiple, complementary, formation controllers for greater robustness. For example, it is possible to have the robot team switch between controllers. Separation-bearing controllers (SBC) rely heavily on the robots sensors [4, 2, 7]. Communications-based controllers rely instead on reliable communications from one robot to the others [3]. If all members of the team dynamically switch between these controllers, together, it would allow the team to compensate for sensor faults using communications, and to compensate for communication faults, by relying on sensors. Such integration is a formidable challenge. The execution of a multi-robot formation controller is often distributed, with each robot individually executing a program associated with its role. For an integrated formation system to be effective, it must have all t...

Citation Context

...aintenance controllers for greater robustness. The robots jointly switch between different controllers, so as to address intermittent failures in sensing or communications. The use of teamwork software in robotics has been reported before [6], but its use for integration of different control schemes is novel. 2. RELATED WORK In formation-maintenance tasks, the objective is to move a group of robots on a desired path, while they maintain their relative position with respect to their peers, according to a desired geometric shape. Many formation maintenance methods have been investigated. (e.g., [2, 8, 5, 7, 3]). These vary in their operating assumptions, their reliance on specific sensors, and their ability to handle different types of faults. Some researchers have examined integration of multiple methods. Fierro et al. [4] analyzed the stability of different controllers, and proposed using manually-constructed control targets to allow up to three robots to switch between alternative schemes, without relying on communications. Fredslund and Mataric [5] describe a method that supplements SBC control by communications, for robustness, thus fusing SBC and communications-based control. Elmaliach and K...

Citation Context

...intaining relative positions with respect to each other (typically describing a specific geometric shape). Indeed, there exists vast literature on various techniques for maintaining formations in a variety of settings. Different controllers have advantages and disadvantages which can be complementary. Little attention has been given to the possibility of integrating together multiple, complementary, formation controllers for greater robustness. For example, it is possible to have the robot team switch between controllers. Separation-bearing controllers (SBC) rely heavily on the robots sensors [4, 2, 7]. Communications-based controllers rely instead on reliable communications from one robot to the others [3]. If all members of the team dynamically switch between these controllers, together, it would allow the team to compensate for sensor faults using communications, and to compensate for communication faults, by relying on sensors. Such integration is a formidable challenge. The execution of a multi-robot formation controller is often distributed, with each robot individually executing a program associated with its role. For an integrated formation system to be effective, it must have all t...

Citation Context

...Conference on Autonomous Agents and Multiagent Systems (AAMAS 2013), Ito, Jonker, Gini, and Shehory (eds.), May, 6–10, 2013, Saint Paul, Minnesota, USA. Copyright c© 2013, International Foundation for Autonomous Agents and Multiagent Systems (www.ifaamas.org). All rights reserved. same controller. The complexity of the communication protocols and the decision-making procedures themselves can be significant. In this abstract, we briefly discuss the use of a teamwork software architecture to automate joint selection of a multi-robot controller, and its execution. Teamwork software (such as BITE [6] or CogniTAO [1]) automates sharing of information and the use of social choice protocols. In particular, it allows synchronous joint selection of controllers for execution. Building on this automation, we describe a formation maintenance system, which integrates together several different formation-maintenance controllers for greater robustness. The robots jointly switch between different controllers, so as to address intermittent failures in sensing or communications. The use of teamwork software in robotics has been reported before [6], but its use for integration of different control schem...

Citation Context

...intaining relative positions with respect to each other (typically describing a specific geometric shape). Indeed, there exists vast literature on various techniques for maintaining formations in a variety of settings. Different controllers have advantages and disadvantages which can be complementary. Little attention has been given to the possibility of integrating together multiple, complementary, formation controllers for greater robustness. For example, it is possible to have the robot team switch between controllers. Separation-bearing controllers (SBC) rely heavily on the robots sensors [4, 2, 7]. Communications-based controllers rely instead on reliable communications from one robot to the others [3]. If all members of the team dynamically switch between these controllers, together, it would allow the team to compensate for sensor faults using communications, and to compensate for communication faults, by relying on sensors. Such integration is a formidable challenge. The execution of a multi-robot formation controller is often distributed, with each robot individually executing a program associated with its role. For an integrated formation system to be effective, it must have all t...

Citation Context

..., there exists vast literature on various techniques for maintaining formations in a variety of settings. Different controllers have advantages and disadvantages which can be complementary. Little attention has been given to the possibility of integrating together multiple, complementary, formation controllers for greater robustness. For example, it is possible to have the robot team switch between controllers. Separation-bearing controllers (SBC) rely heavily on the robots sensors [4, 2, 7]. Communications-based controllers rely instead on reliable communications from one robot to the others [3]. If all members of the team dynamically switch between these controllers, together, it would allow the team to compensate for sensor faults using communications, and to compensate for communication faults, by relying on sensors. Such integration is a formidable challenge. The execution of a multi-robot formation controller is often distributed, with each robot individually executing a program associated with its role. For an integrated formation system to be effective, it must have all the robots in the team switch together from one type of controller to another. Tight coordination is require...