With the rapid development of smart home technology, robotic vacuum cleaners are becoming increasingly popular in household environments. However, the complex and dynamic nature of home settings imposes higher demands on the accuracy and real-time performance of obstacle detection. Existing methods struggle to balance computational efficiency and detection accuracy. To address this challenge, this study proposes an efficient real-time obstacle detection method tailored for robotic vacuum cleaners. The method focuses on constructing a real-time obstacle detection model, optimizing it for lightweight deployment, and implementing it on robotic vacuum cleaners. Innovatively, it combines the YOLO real-time object detection algorithm with model pruning and mobile quantization techniques to enhance detection accuracy and computational efficiency. The proposed method achieves high detection accuracy while improving the operational efficiency of the model on mobile devices. Experimental results demonstrate that the method attains a detection accuracy of 98.5% in complex household environments, with a 35% increase in inference speed and a 50% reduction in model size. This effectively enhances the intelligent perception capabilities of robotic vacuum cleaners and provides robust support for autonomous navigation and environmental sensing in smart home applications.