Webpage-Sinan OGUZ

About me

I am a PhD candidate at Université Libre de Bruxelles and full time researcher at IRIDIA - Artificial Intelligence and Robotics Lab. of ULB.

My research interests are swarm robotics and control theory, more specifically control of networked multi - agent systems. I have access to expert supervision from both of these disciplines, from Prof. Marco Dorigo in swarm robotics, from Prof. Emanuele Garone in control theory.

Research

Swarm robotics research has demonstrated that a large robot swarm can be controlled without any centralized coordinating entity. However, limitations such as development overhead, and the complexity of managing swarm behavior as a whole when programming individual robots, have largely prevented these control approaches from moving to applications.

Therefore, I target the hybridization of centralized and decentralized control, in an attempt to address the limitations of each control approach. To do so, I use an approach based on the existing ‘mergeable nervous systems’ (MNS) technology to control multi-robot formation without the use of global references. The MNS concept combines aspects of centralized and decentralized control, via distributed asymmetric control over a communication graph formed exclusively by self-organization.

The main hypothesis for my Ph.D. is that a swarm robotics system consisting of robots on the ground and drones flying above them will be able to use this MNS technology to realize a system that is more manageable, robust, and trustworthy, without negating the potential parallelism, scalability, and flexibility of swarm robotics systems.

In my PhD research, I also study on formal swarm engineering methodology and analysis techniques for swarm robotics systems. I think that providing a solid theoretical foundation to results obtained empirically could help in understanding the potential capabilities of a swarm system, as well as its inherent limitations. This, in turn, would allow deploying swarms in the real world with more confidence on what can be expected from them.

Swarms Robotics

See more details about Swarm Robotics and Engineering research

Multi-agent Formation Control

See more details about Cooperative Coordination and Formation Control of Multi-agent Systems research

Publications

A simulation-based Development and Verification Architecture for Micro UAV Teams and Swarms

We propose an architecture consisting of five main components which facilitates unmanned aerial vehicle (UAV) system modeling and controller design as well as real time human in the loop applications employing the simulation environments Gazebo, AirSim and V-REP within a single comprehensive framework. This architecture allows joint simulation and testing at both hardware and software layers for multiple vehicle and swarm operations. This architecture has a design that starts with a simple high level controller and evolves into a complex structure in which any changes and tests can be performed with ease.

Decentralized Formation Control of a Swarm of Quadrotor Helicopters

In this paper, we proposed a distributed swarm control mechanism for quadrotor helicopters. The method only re- quires sharing the positions of the swarm members using a simple communication system. Assuming that each quadrotor has single integrator dynamics in the formation control level, we developed a formation control rule and presented the stability analysis of it. We also added other control rules, such as rotation, tracking, and collision avoidance.

Decentralized Formation Control for a Large Scale Swarm of Quadrotor Helicopters

In this letter, we propose a decentralized hybrid control mechanism for a large scale swarm of quadrotor helicopters. First, the mechanism includes a method for finding an appropriate assignment of quadrotors to the locations in the desired formation with collision-free trajectories. Second, it utilizes a formation control approach in which a large number of quadrotors achieves the desired formation within a short period of time without any oscillation. Last, it uses a potential function which provides inter-agent collision avoidance. Our simulation and real world experiments showed that a large number of quadrotors in the random initial locations achieves the desired formation successfully without any collisions.

Formation Control of UAVs and Mobile Robots Using Self-organized Communication Topologies

Formation control in a robot swarm targets the overall swarm shape and relative positions of individual robots during navigation. Existing approaches typically depend on a global reference or a predefined static communication topology. We propose a novel approach without these constraints, by extending the concept of ‘mergeable nervous systems’ to establish distributed asymmetric control via a self-organized wireless communication network.

I am Sinan OGUZ