Motion planning and control of underactuated systems over optimized trajectories

Date

2024-9-5

Author

Koyuncu, Eminalp

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

70
views

47
downloads

Cite This

In this work, we propose an optimal control strategy that is robust and capable of running real-time for nonlinear underactuated systems. Our method combines an optimization-based trajectory planner with the feedback motion planning methods. We combine the local controllers created by the feedback motion planning algorithms to generate a global trajectory with trajectory optimization, taking the underactuation into consideration. We follow the generated trajectory using a global controller. We first generate a Sparse Neighborhood Graph (SNG) in the obstacle-free region of the configuration space. We generate waypoints at each node intersection on the graph, and hierarchical waypoints are identified along the shortest path from start to goal. We then run an optimization algorithm, taking the system dynamics and constraints into account to minimize input effort and generate trajectories between waypoints using a receding horizon optimization strategy. Finally, we use a linear time-varying (LTV) model predictive control (MPC) policy to track the generated trajectory, ensuring constraints are satisfied during system operation. We tested our algorithm on underactuated unmanned surface vehicles (USVs) to drive them in the presence of workspace obstacles, and input and speed constraints. We used 2 different USV models, one implemented in MATLAB and the other is Clearpath Robotics Heron USV on a ROS-Gazebo simulation. We compared our results with previous works, considering real-time performance and robustness. Our work showed superior results regarding all the criterion. However, one drawback of our method is the computational time and power required for the offline planning action.

Subject Keywords

Trajectory optimization, Underactuated systems, Model predictive control

URI

https://hdl.handle.net/11511/111305

Collections

Graduate School of Natural and Applied Sciences, Thesis