Use this URL to cite or link to this record in EThOS:
Title: Sampling-based algorithms for motion planning with temporal logic specifications
Author: Montana, Felipe
ISNI:       0000 0004 7655 4237
Awarding Body: University of Sheffield
Current Institution: University of Sheffield
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
Autonomous mobile robots are machines capable of performing tasks, operating without human intervention. Their presence has increased in applications such as personal assistance, manufacturing, etc. One of the main challenges of controlling a robot autonomously lies in the area of motion planning. This planning needs to consider elements such as system dynamics, uncertainties, dynamic environments, safety and reliability. These requirements have motivated the development of methods that combine control theory and model checking techniques to automatically compute plans that provably guarantee the execution of a given specification. Although model checking has been successfully used to verify discrete systems, its application in autonomous mobile systems presents certain challenges such as: (i) the problem of computing finite models from high-dimensional systems with infinite number of states; (ii) the computation of controllers for systems with kinematic and dynamic constraints; (iii) the computation of robust and reactive controllers to deal with uncertainties and dynamic environments; and (iv) the state explosion problem due to the total number of possible behaviour or states when multiple robots are considered. The challenges presented above are addressed in this thesis. Specifically, the proposed methods in this work are focused on the motion planning of mobile systems based on linear temporal logic and metric interval temporal logic specifications. They are based on sampling methods which are widely used in the point to point motion planning for high-dimensional systems with dynamic constraints. By using these methods and automata-based theory, the solutions in this thesis mitigate the state explosion problem presented in available methods. The main contributions of the thesis are summarised as follows. Chapter 4 develops a new algorithm to find optimal trajectories for deterministic systems with kinematic and differential constraints subject to co-safe temporal logic specifications. Systems with uncertainty in motion and sensing are considered in Chapters 5 and 6. In these chapters, two novel approaches to maximise the probability of completing temporal logic specifications are proposed. Finally, Chapter 7 presents a solution for multi-robot systems subject to cosafe linear temporal logic specifications. All the proposed algorithms are demonstrated with several numerical examples.
Supervisor: Dodd, Tony ; Liu, Jun Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available