Over the past decade, the control research community has given significant
attention to formation control of multiple unmanned vehicles due to a variety of
commercial and defense applications. Consensus-based formation control is considered to
be more robust and reliable when compared to other formation control methods due to
scalability and inherent properties that enable the formation to continue even if one of the
vehicles experiences a failure. In contrast to existing methods on formation control where
the dynamics of the vehicles are neglected, this dissertation in the form of four papers
presents consensus-based formation control of unmanned vehicles-both ground and aerial,
by incorporating the vehicle dynamics.
First, neural networks (NN)-based optimal adaptive consensus-based formation
control over finite horizon is presented for networked mobile robots or agents in the
presence of uncertain robot/agent dynamics and communication. In the second paper, a
hybrid automaton is proposed to control the nonholonomic mobile robots in two discrete
modes: a regulation mode and a formation keeping mode in order to overcome well-known
stabilization problem. The third paper presents the design of a distributed consensus-based
event-triggered formation control of networked mobile robots using NN in the presence of
uncertain robot dynamics to minimize communication. All these papers assume state
availability.
Finally, the fourth paper extends the consensus effort by introducing the
development of a novel nonlinear output feedback NN-based controller for a group of
quadrotor UAVs. |