The University of Southampton
University of Southampton Institutional Repository

Experimentally verified model predictive control of a hover-capable AUV

Experimentally verified model predictive control of a hover-capable AUV
Experimentally verified model predictive control of a hover-capable AUV
This work presents the development of control systems that enable a hover-capable AUV to operate throughout a wide speed range. The Delphin2 AUV was built as part of this project and is used to experimentally verify the prototype control systems. This vehicle is over-actuated with; four through-body tunnel thrusters, four independently-actuated control surfaces and a rear propeller. The large actuator set allows the Delphin2 to operate at low speeds, using the through-body tunnel thrusters, and at high speeds, using the rear propeller and control surfaces. There lies a region between slow and high speed where neither the control surfaces nor tunnel thrusters are operating optimally. To maintain depth stability, both actuator sets are required to operate simultaneously. The model predictive control (MPC) algorithm is used to control the vehicle given its ability to handle multiple inputs and outputs along with system uncertainties. The basis of MPC is a mathematical model of the system to be controlled. Several experiments were conducted with the Delphin2 AUV to acquire the data necessary to develop this model. Bollard pull tests were used to measure thruster performance whilst wind-tunnel and open water experiments provided a measure of the control surfaces, hull and propeller performance. Depth control is the primary focus of this Thesis, however, pitch and surge control are also addressed. Three controllers are developed in this work, of increasing complexity; a depth and pitch controller for low speed operations, a depth and surge velocity controller for medium to high speed operation, and finally, a depth and surge velocity controller for operation from low to high speed operations. All three controllers are multi-input multi-output (MIMO) and use the MPC algorithm. Input constraints are imposed on both the absolute limits and the rate of change limits. Simulations re performed to aid in the design of each controller before it is implemented on the Delphin2 AUV and experimentally verified. The depth and pitch controller, developed for low speed operation, uses the front and rear vertical thrusters as the system inputs. This case demonstrates the implementation of the MPC algorithm and studies the effects of the various tuning parameters. A model sensitivity study is performed, showing that the controller can handle modelling errors of up to ~30%. The controller is experimentally tested and shows excellent performance with zero steady-state errors although there is an undesirably large overshoot of the depth demand. The simulation and experimental results match closely. The depth and surge controller uses the control surfaces and rear propeller as system inputs. Many of the forces and moments within this system are non-linear functions of the vehicles surge velocity. Therefore the standard MPC algorithm, that utilizes just one linearised model, would not be sufficient to capture the system dynamics of the vehicle throughout the full operational envelope. A time-variant MPC (TV-MPC) algorithm is developed and shown in simulation to have excellent performance. The controller did not perform as well when tested experimentally, however, depth regulation of ~0:3 m was achieved. This degradation in performance is due to inaccuracies in the estimation of the vehicles surge velocity. The final controller is also a depth and surge velocity controller, however, it is tasked with maintaining stability through-out the full speed range of the vehicle. All of the system inputs used for depth control are utilised by this controller; the two vertical through-body tunnel thrusters, horizontal control surfaces and the rear propeller. The design of the controller makes use of the TV MPC algorithm. To improve system performance a modification to the controllers cost function, used within the optimisation process, was made to penalise the use of the thrusters at high speeds. This enables the controller to use the thrusters at low speeds, when performing close range inspections, but then as surge velocity increases and the thrusters are no longer required, they are switched off. Both simulation and experimental results show excellent performance, although when the thrusters switch off, the depth control is similar to that of the previous controller due to poor surge velocity estimation.
Steenson, Leo V.
6b4c7fa7-8c25-4a3e-9d45-e0369b578859
Steenson, Leo V.
6b4c7fa7-8c25-4a3e-9d45-e0369b578859
Turnock, Stephen
d6442f5c-d9af-4fdb-8406-7c79a92b26ce

(2013) Experimentally verified model predictive control of a hover-capable AUV. University of Southampton, Faculty of Engineering and the Environment, Doctoral Thesis, 281pp.

Record type: Thesis (Doctoral)

Abstract

This work presents the development of control systems that enable a hover-capable AUV to operate throughout a wide speed range. The Delphin2 AUV was built as part of this project and is used to experimentally verify the prototype control systems. This vehicle is over-actuated with; four through-body tunnel thrusters, four independently-actuated control surfaces and a rear propeller. The large actuator set allows the Delphin2 to operate at low speeds, using the through-body tunnel thrusters, and at high speeds, using the rear propeller and control surfaces. There lies a region between slow and high speed where neither the control surfaces nor tunnel thrusters are operating optimally. To maintain depth stability, both actuator sets are required to operate simultaneously. The model predictive control (MPC) algorithm is used to control the vehicle given its ability to handle multiple inputs and outputs along with system uncertainties. The basis of MPC is a mathematical model of the system to be controlled. Several experiments were conducted with the Delphin2 AUV to acquire the data necessary to develop this model. Bollard pull tests were used to measure thruster performance whilst wind-tunnel and open water experiments provided a measure of the control surfaces, hull and propeller performance. Depth control is the primary focus of this Thesis, however, pitch and surge control are also addressed. Three controllers are developed in this work, of increasing complexity; a depth and pitch controller for low speed operations, a depth and surge velocity controller for medium to high speed operation, and finally, a depth and surge velocity controller for operation from low to high speed operations. All three controllers are multi-input multi-output (MIMO) and use the MPC algorithm. Input constraints are imposed on both the absolute limits and the rate of change limits. Simulations re performed to aid in the design of each controller before it is implemented on the Delphin2 AUV and experimentally verified. The depth and pitch controller, developed for low speed operation, uses the front and rear vertical thrusters as the system inputs. This case demonstrates the implementation of the MPC algorithm and studies the effects of the various tuning parameters. A model sensitivity study is performed, showing that the controller can handle modelling errors of up to ~30%. The controller is experimentally tested and shows excellent performance with zero steady-state errors although there is an undesirably large overshoot of the depth demand. The simulation and experimental results match closely. The depth and surge controller uses the control surfaces and rear propeller as system inputs. Many of the forces and moments within this system are non-linear functions of the vehicles surge velocity. Therefore the standard MPC algorithm, that utilizes just one linearised model, would not be sufficient to capture the system dynamics of the vehicle throughout the full operational envelope. A time-variant MPC (TV-MPC) algorithm is developed and shown in simulation to have excellent performance. The controller did not perform as well when tested experimentally, however, depth regulation of ~0:3 m was achieved. This degradation in performance is due to inaccuracies in the estimation of the vehicles surge velocity. The final controller is also a depth and surge velocity controller, however, it is tasked with maintaining stability through-out the full speed range of the vehicle. All of the system inputs used for depth control are utilised by this controller; the two vertical through-body tunnel thrusters, horizontal control surfaces and the rear propeller. The design of the controller makes use of the TV MPC algorithm. To improve system performance a modification to the controllers cost function, used within the optimisation process, was made to penalise the use of the thrusters at high speeds. This enables the controller to use the thrusters at low speeds, when performing close range inspections, but then as surge velocity increases and the thrusters are no longer required, they are switched off. Both simulation and experimental results show excellent performance, although when the thrusters switch off, the depth control is similar to that of the previous controller due to poor surge velocity estimation.

PDF
thesis_leo_steenson.pdf - Other
Download (34MB)

More information

Published date: 16 June 2013
Organisations: University of Southampton, Civil Maritime & Env. Eng & Sci Unit

Identifiers

Local EPrints ID: 355697
URI: http://eprints.soton.ac.uk/id/eprint/355697
PURE UUID: 45fb594e-e2c9-44ed-8205-cd528e4dd1f1
ORCID for Stephen Turnock: ORCID iD orcid.org/0000-0001-6288-0400

Catalogue record

Date deposited: 11 Nov 2013 16:43
Last modified: 06 Jun 2018 13:14

Export record

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×