National Technical University of Athens (NTUA)
Κατανεμημένος έλεγχος μη γραμμικών πρακτόρων για προβλήματα συμφωνίας
Abstract
dc:descriptionThe thesis addresses the design problem of distributed control laws for nonlinear multi-agent systems. The main idea of this work is the design of suitable transformations, that reformulate various multi-agent consensus problems into regulation problems of the new variables. This enables the use of classical control methodologies to solve the aforementioned consensus problems.The thesis considers the control design for various multi-agent network problems, such as bipartite consensus, optimal consensus and consensus with prescribed performance, as well as the theoretical analysis of the results. Moreover, simulations that illustrate the validity of the theoretical results are presented, based on realistic systems models used in industrial and other practical applications, such as systems of groups of chemical reactors, robotic manipulators and unmanned aerial vehicles.Initially, the current literature in the related research field is thoroughly reviewed. Then, the main contributions of the thesis to the field are presented. Furthermore, the basic theoretical background, necessary for understanding the results that follow, is reviewed.The next part of the thesis addresses the problem of leaderless bipartite consensus with prescribed performance for nonlinear agents. A new methodology is proposed that includes a new Proportional and Integral (PI) transformation, which enables the solution of the aforementioned problem by recasting it into a problem of regulating the new variables with prescribed performance. The communication between the agents is described by signed graphs that have a spanning tree. Using the proposed transformation, the related problem of output consensus with prescribed performance is solved for nonlinear agents in unknown pure-feedback form. The algorithm ensures that all closed-loop signals remain uniformly bounded and guarantees prescribed performance for the bipartite consensus error. Finally, simulations on a group of single-link robotic manipulators are presented, that illustrate the validity of our theoretical results. The thesis then considers the consensus problem of multi-agent systems with communication time-delays. To solve this problem, a Proportional and delayed Integral (PdI) transformation is proposed, that reformulates the cooperative control with communication time-delays into a simpler regulation problem. Following this approach, the problem of leaderless consensus for multi-agent systems with unknown pure-feedback dynamics, communication time-delays and unknown disturbances is then solved for the first time. It is noted that a large class of time-delays is considered, which includes time-varying and distributed delays among others. The proposed algorithm guarantees that all closed-loop signals remain uniformly bounded and that the multi-agent system achieves output consensus with predefined steady-state error. Simulation results on a swarm of unmanned aerial vehicles illustrate the validity of the theoretical analysis. Subsequently, a new methodology that addresses the problem of output optimal consensus for multi-agent systems is presented. A distributed variable transformation (OCPI) is proposed, that recasts the aforementioned problem into a simpler problem of regulating the new variables. The transformation requires only relative output measurements and is independent of the explicit underlying agent dynamics, and therefore is applicable to a large class of system dynamics. Classical control techniques that take into account the underlying agent dynamics can then be employed to regulate the new variables and solve the optimal consensus problem. The methodology is then employed to solve for the first time the problem of output optimal consensus for multi-agent systems with unknown pure-feedback agents subject to unknown disturbances and state time-delays. Moreover, simulations on a group of Continuous Stirred-Tank Reactors (CSTRs) are presented, that validate the theoretical results.Finally, a new distributed control scheme for the formation control with prescribed performance of unmanned aerial vehicles (UAVs) having three degrees of freedom is presented. The information flow between the UAVs is described by a directed graph that includes a spanning tree that has the leader UAV as the root. Only the relative distance information between the UAVs is required to be available to each UAV as indicated by the communication graph. The UAV dynamics are derived from the kinematic equations and the drag coefficients as well as the masses of the UAVs are considered to be unknown. The controllers are designed following a low-complexity backstepping approach. A Lyapunov-type analysis is presented to prove the efficiency of the proposed methodology. Simulation results are presented that validate the theoretical analysis.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Gkesoulis, Athanasios
- Γκεσούλης, Αθανάσιος
Subjects
dc:subject × 28- Πολυπρακτορικά συστήματα
- Μη γραμμικά συστήματα
- Συστήματα ελέγχου
- Αυτόματος έλεγχος
- Κατανεμημένα συστήματα ελέγχου
- Συμφωνία
- Κατανεμημένος έλεγχος συστημάτων
- Διμερής συμφωνία
- Μη επανδρωμένα οχήματα
- Βέλτιστη συμφωνία
- Χρονικές υστερήσεις
- Multi-agent systems
- Nonlinear systems
- Control systems
- Automatic control
- Distributed control systems
- Consensus
- Control of distributed systems
- Bipartite consensus
- Unmanned vehicles
- Optimal consensus
- Time delays
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Αυτοματισμοί και Συστήματα ελέγχου
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Automation and control systems
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/51906