Abstract
The emergence of mobile robotics, particularly in the automotive industry,introduces a promising era of enriched user experiences and adept handling ofcomplex navigation challenges. The realization of these advancementsnecessitates a focused technological effort and the successful execution ofnumerous intricate tasks, particularly in the critical domain of SimultaneousLocalization and Mapping (SLAM). Various artificial intelligence (AI)methodologies, such as deep learning and reinforcement learning, present viablesolutions to address the challenges in SLAM. This study specifically exploresthe application of reinforcement learning in the context of SLAM. By enablingthe agent (the robot) to iteratively interact with and receive feedback fromits environment, reinforcement learning facilitates the acquisition ofnavigation and mapping skills, thereby enhancing the robot's decision-makingcapabilities. This approach offers several advantages, including improvednavigation proficiency, increased resilience, reduced dependence on sensorprecision, and refinement of the decision-making process. The findings of thisstudy, which provide an overview of reinforcement learning's utilization inSLAM, reveal significant advancements in the field. The investigation alsohighlights the evolution and innovative integration of these techniques.