Robust nonlinear model predictive control for quadrotor collision avoidance with interacting multiple model-based obstacle motion prediction

Date

2025-8

Author

Acar, Beyza

Metadata

Show full item record

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

Item Usage Stats

545
views

0
downloads

Cite This

As unmanned aerial vehicles (UAVs) are deployed more frequently in complex environments, the need for safe and reliable navigation near obstacles becomes increasingly important. This thesis presents a nonlinear model predictive control (NMPC) framework for collision avoidance of quadrotors operating in environments with static and dynamic obstacles. NMPC is used as the main control strategy to handle system dynamics, input limits, and obstacle avoidance constraints in a unified optimization problem. To predict obstacle motion over the prediction horizon, an Interacting Multiple Model Kalman Filter (IMM-KF) is employed. The IMM-KF runs multiple motion models in parallel and combines their outputs to adaptively estimate future positions of moving and stationary obstacles. These predictions are directly integrated into the NMPC formulation, allowing the controller to anticipate potential collisions. In addition, an uncertainty-aware repulsive cost term is included in the NMPC objective function. This term adjusts the strength of avoidance based on the confidence of the obstacle predictions, encouraging early and smooth maneuvers when predictions are reliable, even before the hard constraint is activated. The proposed approach is tested under various reference trajectories, model mismatch and different levels of sensor noise. Simulation results show that the framework maintains good trajectory tracking performance and ensures safe navigation without collisions or violations of system constraints. Thanks to the motion prediction, control inputs near obstacles transition more smoothly, avoiding abrupt changes and enabling safer maneuvers.

Subject Keywords

Trajectory tracking, Collision avoidance, Interacting multiple model Kalman filter, Motion prediction, Nonlinear model predictive control

URI

https://hdl.handle.net/11511/115683

Collections

Graduate School of Natural and Applied Sciences, Thesis