### Abstract

This paper addresses the problem to design a unified control and a fault detection system. The proposed control scheme combines a modified inverse problem technique and the idea of repetitive control. The joint torques are computed in each time instant from the unbiased, minimum variance estimate of the joint coordinates and velocities based on the `exact' robot dynamic model and the previous observations. The estimations are generated by the discrete Kalman filter. The idea of the control system's refinement is to use the optimally estimated trajectory obtained at the previous trial as a better reference trajectory for linearization in each working cycle. The problem of fault detection and isolation is formulated as a problem in Hypothesis Testing by regarding the normal operation of the robot manipulator as the null hypothesis. The error signal is defined as the difference between the actual robot manipulator output and the expected output based on the stochastic system model and the previous output data and generated by the discrete Kalman filter.

Original language | English |
---|---|

Title of host publication | 1993 Int Conf Intell Rob Syst |

Editors | Anon |

Publisher | Publ by IEEE |

Pages | 768-772 |

Number of pages | 5 |

ISBN (Print) | 0780308239 |

Publication status | Published - 1993 |

Event | 1993 International Conference on Intelligent Robots and Systems. Part 2 (of 3) - Yokohama, Jap Duration: Jul 26 1993 → Jul 30 1993 |

### Other

Other | 1993 International Conference on Intelligent Robots and Systems. Part 2 (of 3) |
---|---|

City | Yokohama, Jap |

Period | 7/26/93 → 7/30/93 |

### Fingerprint

### ASJC Scopus subject areas

- Engineering(all)

### Cite this

*1993 Int Conf Intell Rob Syst*(pp. 768-772). Publ by IEEE.

**Repetitive trials based robot control and fault detection system.** / Rudas, I.; Ori, Istvan; Toth, Akos.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

*1993 Int Conf Intell Rob Syst.*Publ by IEEE, pp. 768-772, 1993 International Conference on Intelligent Robots and Systems. Part 2 (of 3), Yokohama, Jap, 7/26/93.

}

TY - GEN

T1 - Repetitive trials based robot control and fault detection system

AU - Rudas, I.

AU - Ori, Istvan

AU - Toth, Akos

PY - 1993

Y1 - 1993

N2 - This paper addresses the problem to design a unified control and a fault detection system. The proposed control scheme combines a modified inverse problem technique and the idea of repetitive control. The joint torques are computed in each time instant from the unbiased, minimum variance estimate of the joint coordinates and velocities based on the `exact' robot dynamic model and the previous observations. The estimations are generated by the discrete Kalman filter. The idea of the control system's refinement is to use the optimally estimated trajectory obtained at the previous trial as a better reference trajectory for linearization in each working cycle. The problem of fault detection and isolation is formulated as a problem in Hypothesis Testing by regarding the normal operation of the robot manipulator as the null hypothesis. The error signal is defined as the difference between the actual robot manipulator output and the expected output based on the stochastic system model and the previous output data and generated by the discrete Kalman filter.

AB - This paper addresses the problem to design a unified control and a fault detection system. The proposed control scheme combines a modified inverse problem technique and the idea of repetitive control. The joint torques are computed in each time instant from the unbiased, minimum variance estimate of the joint coordinates and velocities based on the `exact' robot dynamic model and the previous observations. The estimations are generated by the discrete Kalman filter. The idea of the control system's refinement is to use the optimally estimated trajectory obtained at the previous trial as a better reference trajectory for linearization in each working cycle. The problem of fault detection and isolation is formulated as a problem in Hypothesis Testing by regarding the normal operation of the robot manipulator as the null hypothesis. The error signal is defined as the difference between the actual robot manipulator output and the expected output based on the stochastic system model and the previous output data and generated by the discrete Kalman filter.

UR - http://www.scopus.com/inward/record.url?scp=0027799843&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0027799843&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0027799843

SN - 0780308239

SP - 768

EP - 772

BT - 1993 Int Conf Intell Rob Syst

A2 - Anon, null

PB - Publ by IEEE

ER -