Abstract:
The expanding potential of multi-agent systems (MASs) in different robotic applications requires high-level performance, which can be achieved by maintaining
satisfactory collaboration between robots. Such collaboration requires exchanging
sensory data and commands that result in a large amount of data being available,
and brings about the problem of excessive bandwidth requirement, which often overwhelms available communication resources.
In this thesis, the aim is to control a teleoperated swarm of robots that is supposed
to maintain a desired formation and accomplish a given set of tasks without colliding with obstacles, while controlling its bandwidth consumption. For this, two
schemes of motion control and bandwidth allocation are combined and implemented
on networked control systems. The first scheme deploys semi-autonomous formation control based on the Artificial Potential Fields (APF) algorithm, which allows
each individual robot to avoid local obstacles autonomously and tries to maintain
a desired formation with its neighbors, while the operator is in charge of high-level
control only. The second scheme works on distributing available bandwidth among
the existing communication channels in the network, dynamically, based on various
factors that represent changes occurring in the swarm’s internal state and its environment. The aim is to dynamically allocate bandwidth in a manner that improves
the swarm performance and limits its bandwidth consumption. The proposed combined scheme is tested on unmanned aerial vehicles (UAVs) in the ROS-Gazebo
simulator and on PiCar-S mobile robots in real experiments. The obtained simulation and experimental results under different scenarios demonstrate the advantage
of the proposed algorithm over static bandwidth allocation schemes.
