Year: 2025
Type: conference-paper
Abstract: Developing and testing navigation controllers for mobile robots often involves significant challenges related to onboard sensor limitations and the complexities of sensor fusion. Inaccurate or noisy localization data can obscure the true performance of a control algorithm, and sensor integration represents a time-consuming aspect of robotics projects. This paper presents a framework utilizing an external motion capture (MoCap) system integrated with the Robot Operating System (ROS) to provide high-fidelity, ground-truth localization for a Clearpath Husky robot. We detail the system architecture, including custom ROS tools for waypoint collection and navigation. By decoupling controller development from onboard sensor challenges, this framework enables rapid prototyping and comparative analysis. We demonstrate its utility by evaluating proportional (P) and proportional-derivative (PD) waypoint controllers. While initial PD performance was poor, the framework facilitated rapid tuning, resulting in the PD controller significantly outperforming the P controller in accuracy, albeit with longer execution time. This highlights the framework’s effectiveness in efficiently assessing and iterating on controller designs. Future plans include implementing more complex algorithms (e.g., the ROS navigation stack, ROS 2 implementation) and integrating onboard sensors (Stereo Cameras, LiDAR, GPS), guided by the baseline performance established using this MoCap system.
DOI: 10.5038/vdnn8971
Authors: Daniel Granda, Vladimir Gunichev, Caleb Sanchez, James Vanzwieten, Yufei Tang, Pratik Mukherjee
Institution: Florida Atlantic University
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4411946525