Modelling and Control of an Autonomous Two-Wheeled Vehicle
Alen Turnwald
Cite this publication as
Alen Turnwald, Modelling and Control of an Autonomous Two-Wheeled Vehicle (2020), Logos Verlag, Berlin, ISBN: 9783832586409
41
accesses
accesses
2
quotes
quotes
Descripción / Abstract
With respect to the future urban mobility, modern electrical bicycles, advanced motorcycles and innovative two-wheeled vehicles are arresting enormous amount of attention. Especially, model-based control and optimal trajectory planning for such vehicles are important to the research and development of the future. Therefore, a reliable and yet usable vehicle model as well as a systematic approach to motion control for two-wheeled vehicles are essential, to which this work makes a contribution.
Currently available two-wheeled vehicle models are mostly either too complex to be used for a systematic control synthesis, or too simple such that the physical behaviour of the vehicle is no more represented. In this thesis, a unifying approach to modelling and control for autonomous two-wheeled vehicles is presented. The resulting model is generally valid and physically detailed enough to represent the characteristic dynamical behaviour such as the self-stability. At the same time, it is suited to a systematic control synthesis. Furthermore, the systematic extenddability, for instance by a rider, is demonstrated. The model is validated by simulations and by comparison to well-known models from the literature.
The proposed vehicle model is derived in the Lagrangian and Hamiltonian framework and used for model-based optimal trajectory planning. Furthermore, a passivity-based trajectory tracking controller is designed based on the resulting port-Hamiltonian system using the so-called generalised canonical transformations. Such a controller is physically interpretable and robust against parameter uncertainties. To this end, existing approaches of passivity-based controller design are extended and adjusted for two-wheeled vehicles.
Finally, a prototype two-wheeled vehicle is introduced which is used for experimental validation of the model and to demonstrate motion control algorithms for autonomous two-wheeled vehicles.
Currently available two-wheeled vehicle models are mostly either too complex to be used for a systematic control synthesis, or too simple such that the physical behaviour of the vehicle is no more represented. In this thesis, a unifying approach to modelling and control for autonomous two-wheeled vehicles is presented. The resulting model is generally valid and physically detailed enough to represent the characteristic dynamical behaviour such as the self-stability. At the same time, it is suited to a systematic control synthesis. Furthermore, the systematic extenddability, for instance by a rider, is demonstrated. The model is validated by simulations and by comparison to well-known models from the literature.
The proposed vehicle model is derived in the Lagrangian and Hamiltonian framework and used for model-based optimal trajectory planning. Furthermore, a passivity-based trajectory tracking controller is designed based on the resulting port-Hamiltonian system using the so-called generalised canonical transformations. Such a controller is physically interpretable and robust against parameter uncertainties. To this end, existing approaches of passivity-based controller design are extended and adjusted for two-wheeled vehicles.
Finally, a prototype two-wheeled vehicle is introduced which is used for experimental validation of the model and to demonstrate motion control algorithms for autonomous two-wheeled vehicles.
Índice
- BEGINN
- Notation
- 1 Introduction
- 1.1 Motivation
- 1.2 Objective and contribution of this work
- 1.3 Published content
- 2 Preliminaries
- 2.1 General notations and elementary denitions
- 2.2 Constrained mechanics and nonholonomic systems
- 3 Physics of two-wheeled vehicles
- 3.1 Steering behaviour
- 3.2 Self-stability
- 3.3 The linear benchmark model of bicycles
- 4 Lagrangian and Hamiltonian modelling of two-wheeled vehicles
- 4.1 Existing models for two-wheeled vehicles
- 4.2 General model derivation for nonholonomic vehicles using the metric tensor
- 4.3 Derivation of a planar single-track vehicle model
- 4.4 Derivation of the bicycle model: A 2-body approach
- 4.5 Simulations and model validation
- 4.6 Model simplication and extension
- 5 Motion planning for autonomous two-wheeled vehicles
- 5.1 Introduction to trajectory planning for autonomous vehicles
- 5.2 Trajectory planning for a two-wheeled vehicle by direct collocation
- 6 Controller synthesis for autonomous two-wheeled vehicles
- 6.1 Passivity-based control
- 6.2 Trajectory tracking control using GCT
- 6.3 Trajectory tracking control design for two-wheeled vehicles
- 7 A prototype two-wheeled vehicle for experimental validations
- 7.1 Overall system architecture
- 7.2 Actuation system
- 7.3 Perception System
- 7.4 User Control Unit
- 7.5 Software structure
- 8 Experimental results
- 8.1 Experiments for validation of vehicle motion control
- 8.2 Experiments for validation of the vehicle model
- 9 Conclusion and future work
- Bibliography
- Curriculum Vitae