Distributed control of robotic networks

摘要

The field of robotics is evolving from single monolithic robots to teams of small but interconnected robots that achieve global objectives using local coordination. Coordinated missions for such teams of mobile robots include coordinated estimation, surveillance, and coverage, coordinated satellite alignment, as well as distributed placement and assignment in creating desirable team formations. The fundamental challenge in such problems is the design of local rules, such as distributed controllers and estimators, which by local coordination give rise to the desired global objectives. In this thesis, we present the first distributed, scalable, and verifiable algorithm that achieves dynamic robot placement and assignment using local coordination rules. This is achieved using a combination of multi-destination potential fields and assignment coordination protocols, and results in a dynamic network of hybrid robots that is able to reach any desired assignment or formation. We then address the problem of maintaining connectivity in a robotic network, where the robot nodes are mobile. Our solution to this problem is not only the first distributed solution to this problem, but has been both theoretically and experimentally verified. A novel application of the proposed connectivity control algorithm is in multi-robot flocking, where it results in the first flocking algorithm where network connectivity is no longer an assumption but a control objective instead. Finally, we discuss an extension of the above algorithm to the problem of distributed topology control of ad-hoc sensor networks. Such networks typically consist of sensors that due to power constraints can switch between active and sleep operation modes. The control objective investigated in this thesis is how to regulate the switching process so that communication between any pair of active sensors is always guaranteed. Many of the problems addressed in this thesis have recently attracted considerable attention and are interesting in their own right. Our goal is not only to propose efficient solutions to these particular problems, but also to provide guidelines and new ideas, in the interface of engineering and pure sciences, that can result in more systematic approaches in controlling robotic networks.