Autonomous car racing competitions have become widespread across many countries and events, taking inspiration from renowned motorsports like Formula 1 and IndyCar, among others. Similar to how traditional motorsports have fueled advancements in automotive technology, researchers and students are now anticipating similar contributions from autonomous car racing competitions. Betz et al. have provided a comprehensive review of the current landscape of autonomous car racing in [1] . These competitions feature a diverse range of robot cars, varying in scale from 1:10 to full-scale (refer to Table 1 for specifics). The size differences also extend to the sensory and computing systems employed. While many of these events are tailored for educational purposes, targeting students from high school to postgraduate programs, it is noted that smaller-scale hardware systems may not necessarily utilize the same sensory and computing systems as their real-scale counterparts. This, however, does not imply that the technical challenges in algorithm and program development are any less demanding than those encountered in full-scale or near full-scale cases. The variations in sensory, computing, and hardware resources due to the size differences may give rise to distinct technical problems not encountered by autonomous cars navigating real-world road environments. Nevertheless, fundamental issues, such as planning, motion control, and real-time recognition with onboard systems are common for both scales.
Location estimation for waypoint-based autonomous driving robots primarily relies on SLAM (Simultaneous Localization and Mapping) and GNSS (Global Navigation Satellite System) technologies. GNSS performs well in open areas but struggles with signal issues, while SLAM offers high accuracy in feature-rich environments but faces challenges in open spaces. This study integrates SLAM and GNSS, enabling seamless switching between systems when one encounters difficulties. The combined system ensures consistency by aligning coordinate systems and directional data. Experiments in indoor, forest, and open environments demonstrate that the switching algorithm enables stable autonomous driving in diverse conditions, effectively leveraging the strengths of both systems.
1 September 2023 in Journal of Korea Robotics Society