Analysis of inverse simulation algorithms with an application to planetary rover guidance and control

Flessa, Thaleia (2021) Analysis of inverse simulation algorithms with an application to planetary rover guidance and control. PhD thesis, University of Glasgow.

Abstract

Rover exploration is a contributing factor to driving the relevant research forward on guidance, navigation, and control (GNC). Yet, there is a need for incorporating the dynamic model into the controller for increased accuracy. Methods that use the model are limited by issues such as linearity, systems affine in the control, number of inputs and outputs. Inverse Simulation is a more general approach that uses a mathematical model and a numerical scheme to calculate the control inputs necessary to produce a desired response defined using the output variables. This thesis develops the Inverse Simulation algorithm for a general state space model and utilises a numerical Newton-Raphson scheme to converge to the inputs using two approaches: The Differentiation method converges based on the state and output equations. The Integration method converges based on whether the output matches the desired and is suitable for grey or black-box models. The thesis offers extensive insights into the requirements and application of Inverse Simulation and the performance parameters. Attention is given to how the inputs and outputs affect the Jacobian formulation and ensure an efficient solution. The linear case and the relationship with feedback linearisation are examined. Examples are given using simple mechanical systems and an example is also given as to how Inverse Simulation can be used for determining system input disturbances. Inverse Simulation is applied for the first time for guidance and control of a fourwheeled, differentially driven rover. The desired output is the time history of the desired trajectory and is used to produce the required control inputs. The control inputs are nominal and are applied to the rover without additional correction. Using insights from the system’s physics and actuation, the Differentiation and Integration schemes are developed based on the general method presented in this thesis. The novel Differentiation scheme employs a non-square Jacobian. The method provides very accurate position and orientation control of the rover while considering the limitations of the model used. Finally, the application of Inverse Simulation to the rover is supported by a review of current designs that resulted in a rover taxonomy.

• Item Type: Thesis (PhD)
• Colleges/Schools: College of Science and Engineering > School of Engineering
• Supervisor's Name: McGookin, Dr. Euan and Thomson, Dr. Douglas
• Date of Award: 2021
• Depositing User: Theses Team
• Unique ID: glathesis:2021-82863
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 12 May 2022 14:54
• Last Modified: 12 May 2022 14:59
• Thesis DOI: 10.5525/gla.thesis.82863
• URI: https://theses.gla.ac.uk/id/eprint/82863

Related URLs

• Article DOI
• Enlighten Publications Record
• Article DOI
• Article DOI
• Article DOI
• Enlighten Publications Record
• Enlighten Publications Record