One of the most challenging tasks in robotics is the manipulation of DLOs such as cables and wires. This thesis focuses on SDLOs manipulation, which are DLOs having one or more rigid parts, such as connectors. Two common tasks are addressed: assembly and disassembly of cable harness, which involve connector detection, its grasping and connection or disconnection operations. In this work the first part of the pipeline is studied, and a novel cable tracing skill that allows to follow its profile and find the connector without the knowledge of SDLOs material properties or shape is presented. The cable is held by a fixture, whose pose is known by the robot. There, the robotic arm can grasp the cable and start sliding. Small displacements are performed to follow the contour and an admittance controller is used to make last joint compliant and to align motion direction with cable orientation. Once the extremity is reached, another robotic arm grasps the connector. Simple fingertips are exploited, which can be further customized to increase the number of operations performed. The methodology implemented is general and robust, working for different types of cables and configurations. Skill effectiveness was assessed using two cables different in mechanical properties and diameter and by testing it in different configurations, with connected or unconnected end. In all cases, the goal was achieved, meaning that the connector was successfully found. The skill was also exploited into a complex scenario: the collaborative assembly of a motorbike brake, improving the previous solution by implementing the cable tracing skill.
Nel campo dell’automazione, una delle sfide più difficili da affrontare è la manipolazione di DLOs, ovvero cavi e fili. Questa tesi si concentra sulla manipolazione di SDLOs, cioè DLOs che presentano una o più parti rigide come connettori. Due semplici compiti sono stati affrontati: l'assemblaggio e il disassemblaggio di un cablaggio, che includono il riconoscimento il connettore, successivamente la presa e, infine, le operazioni di connessione o disconnessione. In questo lavoro viene studiata la prima parte della pipeline, proponendo una nuova skill di cable tracing che permette di seguire il profilo del SDLO e trovarne il connettore, senza conoscerne materiale o forma. Il cavo è sorretto da una dima, la cui posa è nota al robot. Qui, il robot può afferrare il cavo e cominciare a scorrerlo. Piccoli spostamenti sono eseguiti per seguirne il profilo, mentre un controllore in ammettenza è utilizzato per rendere cedevole l’ultimo giunto e allineare la direzione del movimento con l’orientamento del cavo. Raggiunta l’estremità, un altro braccio robotico afferra il connettore. Semplici dita sono utilizzate, le quali possono essere ulteriormente personalizzate per aumentare le operazioni possibili. La metodologia implementata è generale e robusta, funziona cioè con diversi tipi di cavi in diverse configurazioni. L'efficacia della skill è stata valutata utilizzando cavi di diverso materiale e diametro e in diverse configurazioni, con l'estremità connessa o meno. In tutti i casi, l’obbiettivo è stato raggiunto e l’estremità correttamente trovata. La skill è stata usata anche in uno scenario complesso, ovvero l'assemblaggio collaborativo del freno di una moto, migliorando la soluzione precedente aggiungendo la skill di cable tracing.
Robust cable tracing skill for SDLOs connectors pose detection and manipulation
BUTTAZZONI, SIMONE
2020/2021
Abstract
One of the most challenging tasks in robotics is the manipulation of DLOs such as cables and wires. This thesis focuses on SDLOs manipulation, which are DLOs having one or more rigid parts, such as connectors. Two common tasks are addressed: assembly and disassembly of cable harness, which involve connector detection, its grasping and connection or disconnection operations. In this work the first part of the pipeline is studied, and a novel cable tracing skill that allows to follow its profile and find the connector without the knowledge of SDLOs material properties or shape is presented. The cable is held by a fixture, whose pose is known by the robot. There, the robotic arm can grasp the cable and start sliding. Small displacements are performed to follow the contour and an admittance controller is used to make last joint compliant and to align motion direction with cable orientation. Once the extremity is reached, another robotic arm grasps the connector. Simple fingertips are exploited, which can be further customized to increase the number of operations performed. The methodology implemented is general and robust, working for different types of cables and configurations. Skill effectiveness was assessed using two cables different in mechanical properties and diameter and by testing it in different configurations, with connected or unconnected end. In all cases, the goal was achieved, meaning that the connector was successfully found. The skill was also exploited into a complex scenario: the collaborative assembly of a motorbike brake, improving the previous solution by implementing the cable tracing skill.| File | Dimensione | Formato | |
|---|---|---|---|
|
2021_12_Buttazzoni.pdf
Open Access dal 29/11/2022
Dimensione
64.47 MB
Formato
Adobe PDF
|
64.47 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/10589/182483