Object Tracking-Based “Follow-Me” for Optionally Unmanned Ground Vehicles

Abstract

The paper presents the implementation of autonomously tracking and following guide simplified solution based on common sensors designed for the navigation of Unmanned Ground Vehicles (UGVs). The study investigates the effectiveness of employing a vision system to extract valuable information from a tracked object identified by a selected marker. Limitations associated with vision sensors and the impact of atmospheric conditions are highlighted in the study, emphasising constraints when relying solely on this sensor type. To address these challenges, the paper emphasizes the necessity of incorporating additional spatial sensors providing precise distance data while maintaining resilience to external conditions. The proposed simplified approach to data fusion, integrating vision systems and data from laser scanners, was validated in real operation environment. The results confirm the usefulness of descried approach for guide tracking navigation of unmanned platforms, highlighting its reliability in adverse weather conditions. Potential directions for system improvement are also addressed. This research makes a contribution to advancing autonomous vehicles capabilities, with a particular focus on improving performance in tracking scenarios.