Nowadays, robotics is well established in the industrial world, and it constitutes the key enabler for mass production in modern societies. The current challenge in the robotics community is to endow robots with the capabilities to reliably operate in an unstructured environment, performing a variety of tasks at close contact with humans. One of the additional requirements robots may have to satisfy is the increased mobility with respect to their industrial counterparts. Indeed, robots must be able to move in the environment so as to accomplish their objectives. iCub is a state-of-the-art humanoid robot which has only recently started to balance on its feet. While the current balancing controller has proved successful in various scenarios, it still misses the capability to properly react to strong pushes by taking steps. This thesis goes in this direction by proposing and implementing a control strategy based on simple models available in the literature. As a peculiar characteristic, this control strategy builds up over the already available balancing controller implemented on iCub, thus extending its capabilities to react to external disturbances with the added possibility to take a step while retaining the advantages of torque control for the interaction with the environment. Experiments in the Gazebo simulator and on the iCub humanoid robot validate the proposed strategy. As an additional contribution, the adoption of steps for balancing strategies has been approached by mean of a Model Predictive Controller. A distinctiveness of this second strategy is the possibility to deal directly with intrinsic hybrid dynamics of the system taken into consideration. The effectiveness of this approach has been proved in simulation.
Oggigiorno, la robotica è un settore ben consolidato nel mondo industriale, costituendo la chiave di volta per la produzione di massa nelle società moderne. L’attuale sfida nella comunità robotica consiste nel dotare i robot della capacità di operare in maniera affidabile in ambienti dalle proprietà che possono essergli sconosciute, portando a termine operazioni a stretto contatto con esseri umani. Un ulteriore requisito che un robot potrebbe dover soddisfare, consiste nel possedere una maggiore mobilità rispetto alla sua controparte industriale. In particolare, un robot dovrebbe essere in grado di muoversi liberamente nell’ambiente che lo circonda, portando a termine i propri obiettivi. iCub è un robot umanoide all’avanguardia che solo recentemente ha acquisito la capacità di bilanciarsi sui suoi piedi. Nonostante l’odierno controllore di bilanciamento abbia dimostrato la propria efficacia in diversi scenari, non possiede la capacità di far effettuare al robot uno o più passi a fronte di spinte di una certa intensità. Questa tesi propone un avanzamento in quella direzione, presentando e implementando una strategia di controllo basata su modelli semplificati disponibili in letteratura. Come caratteristica peculiare, questa strategia di controllo si pone al di sopra del controllore di bilanciamento già disponibile e implementato su iCub, estendendo quindi le sue capacità di reagire a disturbi con la possibilità di effettuare un passo, mantenendo tutti i vantaggi del controllo in coppia per l’interazione con l’ambiente. La strategia proposta è validata sia in simulazione tramite Gazebo sia con il robot umanoide iCub. Come ulteriore contributo, l’adozione di passi come strategia di recupero del bilanciamento è stata trattata per mezzo di un controllore “Model Predictive”. La peculiarità di questa seconda strategia è la possibilità di poter considerare la dinamica ibrida intrinseca al sistema preso in considerazione. L’efficacia di questo approccio è dimostrata in simulazione.
Design of balance recovering strategies for a walking humanoid robot. Testing and validation on the iCub platform
DAFARRA, STEFANO
2015/2016
Abstract
Nowadays, robotics is well established in the industrial world, and it constitutes the key enabler for mass production in modern societies. The current challenge in the robotics community is to endow robots with the capabilities to reliably operate in an unstructured environment, performing a variety of tasks at close contact with humans. One of the additional requirements robots may have to satisfy is the increased mobility with respect to their industrial counterparts. Indeed, robots must be able to move in the environment so as to accomplish their objectives. iCub is a state-of-the-art humanoid robot which has only recently started to balance on its feet. While the current balancing controller has proved successful in various scenarios, it still misses the capability to properly react to strong pushes by taking steps. This thesis goes in this direction by proposing and implementing a control strategy based on simple models available in the literature. As a peculiar characteristic, this control strategy builds up over the already available balancing controller implemented on iCub, thus extending its capabilities to react to external disturbances with the added possibility to take a step while retaining the advantages of torque control for the interaction with the environment. Experiments in the Gazebo simulator and on the iCub humanoid robot validate the proposed strategy. As an additional contribution, the adoption of steps for balancing strategies has been approached by mean of a Model Predictive Controller. A distinctiveness of this second strategy is the possibility to deal directly with intrinsic hybrid dynamics of the system taken into consideration. The effectiveness of this approach has been proved in simulation.File | Dimensione | Formato | |
---|---|---|---|
tesi.pdf
solo utenti autorizzati dal 17/09/2017
Descrizione: Testo della tesi
Dimensione
6.97 MB
Formato
Adobe PDF
|
6.97 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/10589/126262