A mobile platform with a manipulator mounted on top of it is an example of two combined robotic devices. During the execution of some tasks it is necessary to establish cooperation between them in order to deal the existing constraints, e.g., the physical connection. A population of several of these combined devices may also be pointed out as an example where cooperation may emerge. Consider, for instance, the transportation of an object of large dimensions by multiple mobile manipulators, acting in cooperation.
For simple robotic devices, i.e., devices with a small number of degrees of freedom (dof), design methodologies for global task controllers are known. Systems with multiple robots may be considered as a single robot with multiple dofs. The immediate extension of these methodologies for robotic devices with a large number of degrees of freedom leads to a rapid increase in the computational complexity of the algorithms, making them useless for real time applications.
This project addresses this last problem by developing an architecture where, instead of dealing with the overall system as a whole, a decoupling methodology supports the cooperation, i.e., the information interchange among the various robots. The architecture to be developed leads to a cooperative behaviour among a set of robots working towards a common goal and is modular, in the sense that it supports, with no major changes, any number of devices. Also, the decision mechanism under the architecture will be carried out at the level of each device, avoiding centralized decision making modules.
The architecture is supported by a separation principle between the controllability properties of each robot, considered as an isolated entity, and those of the global system. Non linear control theory leads a major role in this area.
The addressed problem is relevant because task execution in large workspaces often require robotic mechanisms combining locomotion and manipulation capabilities. In addition, complex tasks or tasks executed in complex environments may determine that more than one robotic device work together, in cooperation, to fulfill the task. This application, points out two major advantages in the use of multiple robotic devices over the single robotic devices: