A Multimodal Perception System for Wheeled Robots Combining Vision and LiDAR

This paper proposes a multimodal perception system that integrates visual sensors and LiDAR to enhance the environmental perception capabilities and navigation accuracy of wheeled robots. In this study, we designed and implemented a fusion algorithm that effectively combines data from cameras and LiDAR, leveraging complementary information to improve understanding of complex environments. The experimental section involved testing in various typical scenarios, including indoor corridors, outdoor parks, and semi-structured roads, to verify the robustness and effectiveness of the system. Results show that compared to single-modal systems, the proposed multimodal perception system performs better in object detection, obstacle recognition, and path planning. Additionally, this research explores the potential application value of the system in future intelligent transportation and home service robots.