Robotic manipulators are mostly employed in industrial environments, where their tasks can be prescribed with little to no uncertainty. This is possible in scenarios where the deployment time of robot workcells is not prohibitive, such as in the automotive industry. In other contexts, however, the time cost of setting up a classical robotic automation workcell is often prohibitive. This is the case with cellphone manufacturing, for example, which is currently mostly executed by human workers. Robotic automation is nevertheless desirable in these human-centric environments, as a robot can automate the most tedious parts of an assembly. To deploy robots in these environments, however, requires an ability to deal with uncertainties and to robustly execute any given task. In this thesis, we discuss two topics related to autonomous robotic manipulation. First, we address parametric uncertainties in manipulation tasks, such as the location of contacts during the execution of an assembly. We propose and experimentally evaluate two methods that rely on force and torque measurements to produce estimates of task related uncertainties: a method for dexterous manipulation under uncertainties which relies on a compliant rotational degree of freedom at the robot's gripper grasp point and exploits contact with an external surface, and a cooperative manipulation system which is able to identify the kinematics of a two degrees of freedom mechanism. Then, we consider redundancies in dual-arm robotic manipulation. Dual-armed robots offer a large degree of redundancy which can be exploited to ensure a more robust task execution. When executing an assembly task, for instance, robots can freely change the location of the assembly in their workspace without affecting the task execution. We discuss methods that explore these types of redundancies in relative motion tasks in the form of asymmetries in their execution. Finally, we approach the converse problem by presenting a system which is able to balance measured forces and torques at ...


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Dual-Arm Robotic Manipulation under Uncertainties and Task-Based Redundancy


    Beteiligte:
    Almeida, Diogo (Autor:in)

    Erscheinungsdatum :

    2019-01-01


    Medientyp :

    Hochschulschrift


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Klassifikation :

    DDC:    629



    Task planning for robotic manipulation in space applications

    Sanderson, A.C. / Peshkin, M.A. / Homen de Mello, L.S. | IEEE | 1988


    Task planning for robotic manipulation in space applications

    Sanderson, A.C. / Peshkin, M.A. / Homen de Mello, L.S. | Tema Archiv | 1988



    Robotic Aubergine Harvesting Using Dual-Arm Manipulation

    Sepúlveda, Delia / Fernández, Roemi / Navas, Eduardo et al. | BASE | 2020

    Freier Zugriff

    Robotic Manipulation under Uncertainty and Limited Dexterity

    Viña Barrientos, Francisco | BASE | 2016

    Freier Zugriff