Multiagent formation control for target tracking and circumnavigation missions
Time: Fri 2017-02-24 11.00
Location: Automatic Control EES, Osquldas väg 10, 10044 Stockholm
Abstract: In this thesis we propose a distributed control algorithm for a formation control of a multiagent system: the aim is to make all the agents of a network localize a target and circumnavigate it, forming a regular polygon during the rotation. In the first part, three different cases are presented: firstly, the scenario in which the target is stationary, at a position known to all the agents of the network , is considered; subsequently, a modified control algorithm is presented for the scenario in which the stationary target position is unknown to the agents, and an estimate law is required to localize it. Finally the most general situation, in which the target is mobile and unknown, is considered. A convergence analysis and simulations in Matlab-Simulink are presented for all the scenarios. In the second part, we present simulations in ROS for the case of dynamical networks, where a number of agents can be added or removed from the multiagent system. Furthermore a time-triggered version of the adopted control law is presented. In the third part, experiments with the Crazyflies 2.0 are shown, related to one of the cases considered in the first part of the presentation.
Subject area: Automatic Control
Respondent: Antonio Boccia
Opponent: Massimiliano Pandolfo
Supervisor: Antonio Adaldo