DATA DRIVEN CONTROL OF AUTONOMUS ROBOTIC PLATFORM WITH SPARSE FEEDBACK MOTION PLANNING

Date

2026-1-22

Author

Adaş, Yiğit

Metadata

Show full item record

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

Item Usage Stats

37
views

0
downloads

Cite This

Safe and robust navigation in dynamic, uncertain environments presents a significant challenge for autonomous mobile robots, especially car-like vehicles with limited maneuverability. While traditional methods often rely on open-loop trajectory gen- eration followed by feedback control, these approaches struggle with real-time per- formance and the complexities of non-holonomic constraints that restrict a car-like vehicle’s instantaneous motion. Recent feedback motion planning strategies utilizing connected obstacle-free regions offer a promising alternative but have been primarily limited to simulation or holonomic systems. This paper integrates a feedback motion planning method, based on those obstacle-free regions, with a Model Predictive Path Integral (MPPI) controller, specifically designed for car-like vehicles. This integra- tion leverages MPPI’s sampling-based approach to achieve efficient, online control that inherently respects the vehicle’s limited maneuverability. We demonstrate the successful application of this integrated approach and on a physical car-like robotic platform, achieving real-time performance without sacrificing control accuracy. Ex- perimental results validate the feasibility and effectiveness of deploying this feedback motion planning technique in real-world autonomous driving scenarios, where com- putational efficiency and robust navigation are paramount.

Subject Keywords

motion planning, control systems, system identification

URI

https://hdl.handle.net/11511/118736

Collections

Graduate School of Natural and Applied Sciences, Thesis