Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MCE-based Direct FTC Method for Dynamic Positioning of Underwater Vehicles with Thruster Redundancy

Ji-Hong Li

TL;DR

This work tackles dynamic positioning of thruster-redundant underwater vehicles under unmatched disturbances by introducing an active model-based fault-tolerant control (FTC) framework that leverages the motion control error (MCE) to form residuals for fault detection and trends in MCE for fault identification. The method eliminates reliance on estimation-error residuals and instead derives fault diagnosis and thrust-loss estimation directly from the trajectory tracking error, using a backstepping controller to achieve uniform ultimate boundedness in the presence of disturbances. A key contribution is the ability to identify up to two simultaneous thruster faults and to estimate thrust losses via an online adaptation mechanism, with a control allocation step that reconfigures the thrust distribution accordingly. Numerical studies on a 3DOF horizontal model of BlueROV2 Heavy ROV in strong currents demonstrate effective fault detection, identification, and thrust-loss estimation, highlighting both robustness and practical applicability while also noting limitations in certain low-sensitivity fault configurations and the potential need for extending to triple-fault cases.

Abstract

This paper presents an active model-based FTC (fault-tolerant control) method for the dynamic positioning of a class of underwater vehicles with thruster redundancy. Compared to the widely used state and parameter estimation methods, this proposed scheme directly utilizes the vehicle's motion control error (MCE) to construct a residual for detecting thruster faults and failures in the steady state of the control system. In the case of thruster fault identification, the most difficult aspect is that the actual control input with thruster faults is unknown. However, through a detailed and precise analyses of MCE variation trends in the case of thruster faults, highly useful information about this unknown control input can be extracted. This characteristic also serves as the foundation for the novel scheme proposed in this paper. As for control reconfiguration, it is straightforward since the thrust losses can be directly estimated as a result of the identification process. Numerical studies with the real world vehicle model are also carried out to demonstrate the effectiveness of the proposed method.

MCE-based Direct FTC Method for Dynamic Positioning of Underwater Vehicles with Thruster Redundancy

TL;DR

This work tackles dynamic positioning of thruster-redundant underwater vehicles under unmatched disturbances by introducing an active model-based fault-tolerant control (FTC) framework that leverages the motion control error (MCE) to form residuals for fault detection and trends in MCE for fault identification. The method eliminates reliance on estimation-error residuals and instead derives fault diagnosis and thrust-loss estimation directly from the trajectory tracking error, using a backstepping controller to achieve uniform ultimate boundedness in the presence of disturbances. A key contribution is the ability to identify up to two simultaneous thruster faults and to estimate thrust losses via an online adaptation mechanism, with a control allocation step that reconfigures the thrust distribution accordingly. Numerical studies on a 3DOF horizontal model of BlueROV2 Heavy ROV in strong currents demonstrate effective fault detection, identification, and thrust-loss estimation, highlighting both robustness and practical applicability while also noting limitations in certain low-sensitivity fault configurations and the potential need for extending to triple-fault cases.

Abstract

This paper presents an active model-based FTC (fault-tolerant control) method for the dynamic positioning of a class of underwater vehicles with thruster redundancy. Compared to the widely used state and parameter estimation methods, this proposed scheme directly utilizes the vehicle's motion control error (MCE) to construct a residual for detecting thruster faults and failures in the steady state of the control system. In the case of thruster fault identification, the most difficult aspect is that the actual control input with thruster faults is unknown. However, through a detailed and precise analyses of MCE variation trends in the case of thruster faults, highly useful information about this unknown control input can be extracted. This characteristic also serves as the foundation for the novel scheme proposed in this paper. As for control reconfiguration, it is straightforward since the thrust losses can be directly estimated as a result of the identification process. Numerical studies with the real world vehicle model are also carried out to demonstrate the effectiveness of the proposed method.

Paper Structure

This paper contains 19 sections, 28 equations, 10 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Methodology
  4. Problem Statement
  5. Vehicle Model
  6. Thruster Fault and Failure Model
  7. Problem Formulation
  8. Motion Control Design
  9. Step 1
  10. Step 2
  11. Fault Detection and Identification
  12. Fault Detection
  13. Fault Identification
  14. Control Allocation
  15. Numerical Studies
  16. ...and 4 more sections

Figures (10)

  • Figure 1: Vehicle configuration in the navigation frame and its thruster configuration in the body-fixed frame.
  • Figure 2: Thruster fault occurrence scenario.
  • Figure 3: Corresponding residual measurement.
  • Figure 4: Corresponding time variations of designed controller $\bm{\tau}_c$ and actual control input $\bm{\tau}$.
  • Figure 5: Vehicle trajectory during the dynamic positioning without thrust loss parameter update.
  • ...and 5 more figures