University of Twente Student Theses


Toward spatial impedance estimation for robotic systems

Graziosi, Damiano (2018) Toward spatial impedance estimation for robotic systems.

[img] PDF
Abstract:In present days, performing complex tasks from a remote site through the use of intermediate mechatronic systems is not an illusion anymore (one can think of the Da Vinci surgical system, for example). That is why telepresence and telemanipulation are topics of research and development within the i-Botics Joint Innovation Center - a collaboration between the University of Twente and TNO. Combining the cognitive ability of the human operator with the robotic capabilities at a distance is the main objective of this cooperation. Robots never get tired, neither bored nor injured. However, they are still far away from understanding a situation through the senses human beings have been developing for thousand of hundreds of years of evolution on Earth. Therefore, improving these sensing capabilities would allow them to be used even more in complicated tasks where those human abilities and “situational awareness” are required, avoiding human presence in hazardous scenarios (one can think of EOD or pipe inspection robots). In a typical teleoperated system, unavoidable latencies in between operator (master) and the end-robot (at the slave site) jeopardise transparency (haptic performance) and endanger stability. However, it is known from previous studies that having available a more accurate model of the remote environment, i.e. where the manipulation takes place, would enhance the overall performance the system. For this purpose, an adaptive impedance law can be implemented on each controller side when force/position measurements are available within the Bilateral Impedance Control architecture. Within a typical teleoperation scheme as in figure 1, this assignment focused on the remote side where, embracing previous implementations, it aimed to enhance the recognition of the real environment. It has been accomplished by means of an on-line estimation of the dynamic properties of the physical surrounding, when a 7 DOF telecontrolled robotic arm interacts with it. Within the same operational scenario and making use of these estimates, the system can be then trained to have available a first rough assessment of those physical properties by the only visual data. Eventually, these attributes can clustered and associated to the corresponding item in the reconstructed virtual world, so as to end up with a “multisensorial picture” of the reality.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:52 mechanical engineering
Programme:Systems and Control MSc (60359)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page