Supervisor: Prof. Ville Kyrki (email@example.com)
Advisor: Dr. Gökhan Alcan (firstname.lastname@example.org)
Keywords: differential dynamic programming, state constraints, input constraints, constrained optimization.
Differential Dynamic Programming (DDP) is one of the most successful trajectory optimization method, in which a large optimization problem is decomposed into smaller optimization sub-problems iteratively. This is the most important advantage of DDPs over collocation type methods. Even though the key advantage of collocation methods over DDPs is their abilities to handle state and control constraints, there exist some recent studies to employ state and/or control constraints in a DDP framework as well.
In this thesis, an extensive investigation of constrained DDP methods will be performed and the major selected ones will be implemented in simulation environment for trajectory optimizations of different robots such as a simple point robot, 2D car-like robot, 3D quadrotor robot and cart-pole system. In this context, the methods will be compared in terms of convergence speed, computational complexity, sensitivity to initializations and parameter selections.
- Related literature review,
- Design of the simulation environment for different robots such as simple point robot, 2D car-like robot, 3D quadrotor robot and cart-pole system,
- Implementation of the major selected constrained DDP methods in these simulation environments,
- Comparisons of the methods in terms of convergence speed, computational complexity, sensitivity to initializations and parameter selections.
Pre-requisites: Python(high), numerical optimization (medium)
Tools: OpenAI Gym, OSQP
Start: Available immediately
- Z. Xie, C. K. Liu and K. Hauser. “Differential Dynamic Programming with Nonlinear Constraints“. IEEE International Conference on Robotics and Automation, pp. 695-702, 2017.
- Y. Aoyama, G. Boutselis, A. Patel and E. A. Theodorou. “Constrained Differential Dynamic Programming Revisited“. arXiv:2005.00985, 2020.