Cooperative control of autonomous multi-agent systems is a research area which is getting significant attention in recent years. Multi-agent systems allow for a broad spectrum of applications and cooperation can increase their flexibility, efficiency and robustness to changes in external constraints and disturbances. Focusing on autonomous vehicles, examples of possible applications of cooperative multi-agent systems include search and rescue missions, autonomous delivery and performing of emergency landings. The purpose of the thesis is to develop and implement a cooperative rendezvous algorithm based on model predictive control and apply it to the problem of autonomous landing in an indoor setting. The agents involved in the maneuver are a quadcopter and a ground carrier. The two agents cooperatively negotiate on the optimal location for the touchdown taking also into account relevant spatial constraints and, if necessary, update that location, also referred to as rendezvous point, in real-time throughout the maneuver. The algorithm is first tested and validated in a simulated environment and finally on the physical system during real-time operations. Additional scenarios are tested in the simulated environment in order to further inspect the potential capabilities of the developed algorithm. Those additional simulations analyse how the algorithm behaves when a constant lateral wind influences the quadcopter; when the controllers operate at a reduced frequency; and when the quadcopter is affected by an external Gaussian disturbance. The developed algorithm proved to be suitable for the purpose, allowing the agents to perform the desired maneuver in a relatively short time and with a high degree of precision.
Il controllo cooperativo di sistemi multi-agente autonomi è un'area di ricerca che sta attraendo significativa attenzione negli ultimi anni. I sistemi multi-agente permettono di effettuare un ampio spettro di operazioni e la cooperazione può accrescere la loro flessibilità, efficienza e robustezza a variazioni dei vincoli e disturbi esterni. Concentrandosi sui veicoli autonomi, esempi di possibili applicazioni di sistemi cooperativi multi-agente includono missioni di search and rescue, consegna autonoma e l'effettuazione di atterragi di emergenza. lo scopo della tesi è di sviluppare e implementare un algoritmo per un rendezvous cooperativo basato su una strategia di controllo predittivo e applicarlo al problema dell'atterraggio autonomo in un ambiente indoor. Gli agenti coinvolti nella manovra sono un quadrirotore e un carrier terrestre. I due agenti negoziano cooperativamente la posizione ottimale per il touchdown tenendo anche in considerazione adeguati vincoli spaziali e, se necessario, aggiornano quest'ultima, anche chiamata punto di rendezvous, in real-time nel corso della manovra. L'algoritmo è inizialmente testato e validato in un ambiente simulato e infine testato sul sistema fisico durante operazioni in real time. Scenari addizionali sono testati nell'ambiente simulato con lo scopo di studiare ulteriormente le potenziali capacità dell'algoritmo sviluppato. Queste ulteriori simulazioni analizzano come l'algoritmo si comporta quando un vento laterale costante influnza il quadrirotore; quando i controllori operano ad una frequenza ridotta; e quando il quadrirotore è influenzato da un disturbo Gaussiano esterno. L'algoritmo sviluppato ha dimostrato di essere adatto allo scopo, permettendo gli agenti di effettuare la manovra desiderata in un tempo relativamente corto e con un alto grado di accuratezza.
Autonomous landing of an UAV on a moving UGV platform using cooperative MPC
Garegnani, Luca
2020/2021
Abstract
Cooperative control of autonomous multi-agent systems is a research area which is getting significant attention in recent years. Multi-agent systems allow for a broad spectrum of applications and cooperation can increase their flexibility, efficiency and robustness to changes in external constraints and disturbances. Focusing on autonomous vehicles, examples of possible applications of cooperative multi-agent systems include search and rescue missions, autonomous delivery and performing of emergency landings. The purpose of the thesis is to develop and implement a cooperative rendezvous algorithm based on model predictive control and apply it to the problem of autonomous landing in an indoor setting. The agents involved in the maneuver are a quadcopter and a ground carrier. The two agents cooperatively negotiate on the optimal location for the touchdown taking also into account relevant spatial constraints and, if necessary, update that location, also referred to as rendezvous point, in real-time throughout the maneuver. The algorithm is first tested and validated in a simulated environment and finally on the physical system during real-time operations. Additional scenarios are tested in the simulated environment in order to further inspect the potential capabilities of the developed algorithm. Those additional simulations analyse how the algorithm behaves when a constant lateral wind influences the quadcopter; when the controllers operate at a reduced frequency; and when the quadcopter is affected by an external Gaussian disturbance. The developed algorithm proved to be suitable for the purpose, allowing the agents to perform the desired maneuver in a relatively short time and with a high degree of precision.| File | Dimensione | Formato | |
|---|---|---|---|
|
Autonomous_Landing_Of_An_UAV_On_A_Moving_UGV_Platoform_Using_Cooperative_MPC.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: ElaboratoDiTesi
Dimensione
9.41 MB
Formato
Adobe PDF
|
9.41 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/179661