Master Thesis

Supervision and Fault Tolerance for Assistive Robotics

Work default illustration


  • Started: 05/02/2019
  • Finished: 15/07/2019


In this Master Thesis, the supervision and control problem of service robots in unknown anthropic domains has been addressed from the Fault Detection and Diagnosis (FDD) framework, presenting a complete Fault-Tolerant scheme able to detect, isolate and compensate the effects of an exogenous force acting on a robotic manipulator. Therefore, a systematic approach has been presented, applied to the TIAGo head subsystem, to obtain a Takagi-Sugeno representation suitable for a Parallel Distributed Controller, with the main advantage of defining the complete behaviour of the system using only its representation at the operational limits. Additionally, the Robust Unknown Input Observer for Takagi-Sugeno Models has been implemented for an incomplete information model scenario, which allows decoupling the given estimation from the effect of exogenous faults, disregarding its behaviour nor eventuality. Finally, a characterization of the real robot actuators has been performed, in order to design the suitable mechanisms for their implementation into the complete Fault-Tolerant scheme.


The work is under the scope of the following projects:

  • MdM: Unit of Excellence María de Maeztu (web)
  • SIMBIOTS: Facilitar una introducció de la robòtica a nous processos i aplicacions dintre de la industria (web)